Developments and Applications of Geomatics: Proceedings of DEVA 2022 (Lecture Notes in Civil Engineering, 450) 9819985676, 9789819985678

Table of contents :
Foreword
Preface
Acknowledgements
Contents
About the Editors
Abbreviations
Analyzing the Potential Application of Low-Cost Digital Image Correlation in Direct Shear Test
1 Introduction
1.1 DIC Applications in Lab Testing of Materials
1.2 Direct Shear Test
2 Materials and Method
2.1 Shear Mold
2.2 2D-DIC Fundamentals
2.3 Experimental Setup
2.4 Validation of Results
3 Results and Discussion
3.1 Noise Floor Analysis
3.2 Accuracy Assessment
4 Conclusions
References
Applications of GIS in Estimating the Probable Maximum Earthquake Magnitude for Amaravati Region, Andhra Pradesh, India
1 Introduction
2 Details of the Study Area
3 Characterization of Seismicity
4 Estimation of Maximum Magnitude (Mmax)
4.1 Gupta (2002) (Magnitude Incremental Method)
4.2 Kijko-Sellevoll-Bayes (KSB) Method
5 Conclusions
References
Assessing the Effect of Land Use Land Cover Change on the Water Quality Index of a River Basin Using GIS and Remote Sensing Techniques
1 Introduction
1.1 Objectives
1.2 Study Area
1.3 Water Quality Index
1.4 LULC Mapping Using GIS and Remote Sensing
2 Methodology
2.1 General
2.2 Overview of Procedure
3 Results and Discussion
3.1 Estimation of WQI
3.2 LULC Supervised Classification
4 Conclusion
4.1 Scope of Future Work
References
Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground Water Levels in Kurukshetra, Haryana
1 Introduction
1.1 General
1.2 Significance of Present Study
1.3 Objective of the Study
2 Study Area
3 Methodology
3.1 General
3.2 IDW Contours in ArcGIS
3.3 Box Plot
4 Result and Discussion
4.1 Groundwater Table Level Average in the Pre- Monsoon Season in 2010
4.2 Groundwater Table Level Average in the Pre- Monsoon Season in 2020
4.3 Groundwater Table Level Average During Post Monsoon Period in 2010
4.4 Groundwater Table Level Average During Post Monsoon Period in 2020
5 Conclusion
References
Assessment of Land Use—Land Cover Changes in District Dehradun (1991–2021)
1 Introduction
1.1 The Study Area
1.2 Need for the Study
2 Data and Methodology
2.1 Data Used
2.2 Methodology
2.3 Accuracy Assessment
3 Results and Discussions
3.1 Problems Faced During Classification
3.2 Classification Accuracy
3.3 Relation of LULC with Population
3.4 Impact of LULC Changes
4 Conclusion
References
Comparison of Streamflow Simulations for Different DEMs
1 Introduction
1.1 Hydrologic Engineering Center-Hydrologic Modeling System (HEC-HMS) Model
1.2 Digital Elevation Model (DEM)
2 Materials and Methods
2.1 Data
2.2 Methodology
2.3 Study Area
3 Results and Discussions
3.1 Results
3.2 Conclusion
References
Comprehensive Analysis of Impact of COVID-19 Lockdown on Air Quality in Andhra Pradesh, India
1 Introduction
2 Methodology
2.1 Study Area and Data Source
2.2 Data Analysis
2.3 HYSPLIT Backward Trajectory Analysis (BTA)
3 Result and Discussion
3.1 Effects of Lockdown on Air Quality
3.2 Meteorology and Air Quality
3.3 Source Identification
4 Conclusion
References
Development of Mobile Application for Assessing Urban Heat Island (UHI) Using Geospatial Techniques a Case Study of Chennai City
1 Introduction
1.1 Study Area
2 Data Sets Used
2.1 Satellite Datasets
2.2 Other Datasets
3 Methodology
3.1 UHI Mapping
3.2 Heat Stress Index (WBGT) Mapping
3.3 Heat Stress-Related Illnesses Mapping
3.4 Development of Mobile Application
4 Results and Discussion
4.1 Spatial Distribution of UHI Hotspots
4.2 Spatial Distribution of WBGT
4.3 Examination of Health Impact
4.4 Interpretation from GIS Analysis
4.5 Information from Mobile Application
5 Conclusion
References
Drones as an Alternate Communication System During Calamities
1 Introduction
2 Preceding Work
3 Mathematical Model
3.1 Logical Settings for the System
3.2 Investigation of Path Deficit Models
3.3 Path Loss Model Equilibrium
4 Devising of Mathematical Problem
4.1 Constraints for Mathematical Problem
5 Methodology
6 Technical Parameters for Simulation
7 Water-Filling Algorithm
8 Results and Discussion
9 Conclusion
References
Drought Analysis of an Area Using Google Earth Engine
1 Introduction
1.1 Drought
1.2 Drought in India
1.3 About Google Earth Engine
2 Literature Review
3 Procedure
3.1 Procedure for Calculation of SPI
3.2 Procedure for Calculation of DSI
4 Database Collection from GEE
5 Results and Discussions
5.1 Jangaon District
5.2 Karimnagar District
5.3 Mahbubnagar District
5.4 Nizamabad District
5.5 Khammam District
6 Summary and Conclusions
References
Effects of Urbanization on Land Use Land Cover of Warangal Region Using RS and GIS
1 Introduction
1.1 Effects of Urbanization on LULC
1.2 Study Area
2 Materials and Methods
3 Results
3.1 LULC Details of Warangal Urban Region
3.2 LULC Details of Warangal Rural Region
3.3 Population Growth of Warangal
3.4 Annual Rainfall of Warangal
4 Conclusions
References
Effect of LULC Changes on Land Surface Temperature
1 Introduction
1.1 Land Surface Temperature
1.2 Urban Heat Island
1.3 Objectives
2 Study Area
2.1 Data Used
2.2 Methodology
2.3 Image Pre-processing
2.4 Image Classification
2.5 Spectral Indices
2.6 LST Estimation from Landsat 7
2.7 LST Estimation from Landsat 8
2.8 NDVI Method for Emissivity Correction Calculation
2.9 Future Prediction of LULC
3 Results and Discussions
3.1 Land Use Land Cover Analysis
3.2 NDVI
3.3 Land Surface Temperature Analysis
3.4 Future Prediction
4 Conclusions
References
Estimation of Aerosol Direct Radiative Forcing in Southern India
1 Introduction
2 Data
3 Methodology
3.1 Calculation of ADRF
4 Results and Discussions
4.1 ADRF at TOA, Surface and in the Atmosphere
5 Conclusions
References
Estimation of Groundwater Potential Zones in Southern Dry Agro-Climatic Area Using Geoinformatics and AHP Technique
1 Introduction
2 Materials and Method
2.1 Thematic Layer Processing
2.2 Analytic Hierarchy Process
3 Results and Discussion
3.1 Assessment of Groundwater Potential Zones
3.2 Validation of Groundwater Potential Zone
4 Conclusion
References
Evaluation and Prediction of Land Use and Land Cover Changes in the Kumaradhara Basin, Western Ghats, India
1 Introduction
2 Materials and Methodology
2.1 Study Area
2.2 Data Used in the Study
2.3 Methodology
3 Results and Discussion
3.1 Comparison of Landcover Datasets
3.2 Accuracy Assessment and Change Analysis of Classified Images
3.3 LULC Change Detection
3.4 Forecasting Future LULC Changes
4 Conclusion
References
Evaluation of Surface Soil Moisture Using Remote Sensing and Field Studies
1 Introduction
2 Study Area
3 Data
3.1 Sentinel-1A
3.2 Sentinel-2A
3.3 Field Data
4 Method
4.1 Image Processing
4.2 Data Collection from Field
5 Analysis
5.1 Field Data Analysis
5.2 Soil Moisture Retrieval Model
6 Results and Discussion
6.1 Surface Soil Roughness and Backscattering Energy
6.2 Vegetation and Backscattering Energy
6.3 Dielectric Constant and Backscattering Energy
6.4 Soil Moisture and Backscattering Energy
6.5 Model Development
6.6 Validation
7 Conclusion
References
Evaluation of the Influence of Land Use and Climate Changes in Runoff Simulation Using Semi-Distributed Hydrological Model
1 Introduction
2 Materials and Methods
2.1 Study Area
2.2 Input Data
2.3 Description of the SWAT Model
2.4 Calibration and Validation
2.5 Separation of the Impact of Climate and LULC Change
3 Results and Discussion
3.1 Sensitivity Analysis, Calibration, and Validation
3.2 Relative Contribution Assessment
4 Conclusions
References
Flood Damage Assessment of a River Basin Using HEC-GeoRAS
1 Introduction
1.1 General Background
1.2 Objectives
1.3 Study Area
2 Literature Review
2.1 General
2.2 Flood Damage Assessment
3 Methodology
3.1 General
3.2 Description of the HEC-GeoRAS Model
3.3 Watershed Delineation of Vamanapuram River Basin
3.4 Preparation of Flood Inundation Map
3.5 Flood Damage Assessment
4 Results and Discussion
4.1 General
4.2 Watershed Delineation of Vamanapuram River Basin
4.3 Preparation of Flood Inundation Map
5 Conclusion
References
Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques
1 Introduction
2 Study Area
3 Methods and Methodology
3.1 Data Used
3.2 Methodology
4 Results and Discussion
5 Conclusions
References
GIS and RS-Based Soil Erosion and Sediment Yield Modelling in Manikpur, Chhattisgarh, India
1 Introduction
2 Study Area
2.1 Data Used
3 Methodology
3.1 Soil Erosion and Sediment Yield Modelling Using Empirical Equations
3.2 Direct Method for the Estimation of Sediment Yield
4 Results and Discussion
5 Conclusions
References
Groundwater Level Trends Over Southern India
1 Introduction
2 Study Area
3 Data and Methods
3.1 Mann–Kendall Test
3.2 Results
4 Conclusions
References
Impact of Climate Change on Streamflow Over Nagavali Basin, India
1 Introduction
2 Materials and Methods
2.1 Study Area
2.2 Datasets
2.3 Model Setup and Methodology
3 Results and Discussions
3.1 Flood Frequency Analysis
3.2 Calibration and Validation
3.3 Climate Change Impact on Precipitation
3.4 Climate Change Impact on Streamflow
4 Conclusions
References
Impervious Surface Area Prediction Using Landsat Satellite Imagery and Open Source GIS Plugin
1 Introduction
2 Study Area
3 Methodology
3.1 Data Collection
3.2 Data Preparation
3.3 Data Processing
4 Results and Discussion
5 Conclusion
References
Influence on Water Characteristics Away from Various Sources of NIT Kurukshetra Using ArcGIS
1 Introduction
2 Study Area
3 Materials and Methods
4 Results and Discussion
5 Conclusion
References
Landslide Hazard Zonation Mapping Using Remote Sensing and GIS in Mountainous Terrain
1 Introduction
2 Objective of the Study Area
3 Description of Study Area and Datasets Used
4 Methodology
4.1 GIS Analysis for Landslide Hazard Zonation Mapping
4.2 Analytic Hierarchy Process (AHP)
5 Results and Discussions
5.1 Thematic Parameters and Their Relationship with Landslide Causes
5.2 Pairwise Comparison-Matrix Development
5.3 Computation of Criterion Weights
5.4 Consistency-Ratio Estimation
6 Landslide Hazard Zonation (LHZ) Map: 53 J/8
6.1 Landslide Inventory Mapping
7 Summary and Conclusions
References
Modeling Daily Streamflow from Idamalayar Catchment Using SWAT
1 Introduction
2 Study Area and Data Used
2.1 Study Area
2.2 Data Used
3 Methodology
3.1 SWAT Model Description
3.2 Calibration and Validation of the Model
3.3 Model Evaluation Using Statistical Performance Measures
4 Results and Discussions
5 Conclusion
References
Modelling the Low Impact Development Alternatives for Rainfall Runoff Reduction
1 Introduction
2 Data
3 Methodology
3.1 Implementation of LID Controls in PCSWMM
4 Runoff Simulation for Different Scenarios Analysis (W.R.T to 2 yr Design Storm and Combination of LIDs)
5 Conclusions
References
Performance Evaluation of Support Vector Machine and Random Forest Techniques for Land Use-Land Cover Classification—A Case Study on a Mili Scale Agricultural Watershed, Tadepalligudem, India
1 Introduction
2 Materials and Methods
2.1 Study Area
2.2 Materials
2.3 Methodology
3 Results and Discussions
4 Conclusions
References
Photogrammetric Survey of an Intertidal Area: A Case Study in NW Spain
1 Introduction
2 Material and Methods
2.1 Equipment and Software Used
2.2 Methodology
3 Results and Discussion
3.1 Flight Execution
3.2 Image Processing
3.3 Bathymetric Survey by Sonar
4 Conclusion
References
Potential Zones Identification to Effectively Exploit Solar and Wind Energy in the State of Assam—A MCDA Approach Using GIS and Remote Sensing
1 Introduction
2 Study Area
3 Literature Review
4 Materials and Methods
4.1 Criteria for Solar Systems Installation
4.2 Criteria for Wind Farm Siting
4.3 Analytic Hierarchy Process (AHP)
5 Results and Concluding Remarks
References
Prediction of Soil Organic Carbon in Unscientific Coal Mining Area Using Landsat Auxiliary Data
1 Introduction
2 Material and Methodology
2.1 Study Area
2.2 Soil Sampling and Process
3 Result and Discussion
3.1 General Statistics
3.2 Relationship Between Indices and SOC
3.3 Spatial Modelling and Prediction of SOC Content
3.4 Validation
4 Conclusion
References
Rainfall Runoff Modeling Using HEC-HMS for Munneru River Basin, India
1 Introduction
2 Study Area
3 Data Used
4 Methodology
4.1 Hec-Hms
5 Results and Discussion
6 Conclusion
References
Spatio-Temporal Surface Urban Heat Island Effect Analysis Over Tiruchirappalli City, India, Using GIS Techniques
1 Introduction
2 Study Area
3 Methodology
4 Results and Discussions
5 Conclusions
References
Simulation of Streamflow and the Assessment of Nutrient Loadings for the Indravati River Basin of India using SWAT
1 Introduction
2 Materials and Methods
2.1 Study Area
2.2 Description of SWAT Model
2.3 Data Sources
2.4 Digital Elevation Model
2.5 LULC Map
2.6 Soil Map
3 Results and Discussion
3.1 Sub-basins and HRU Creation
3.2 Hydrologic Cycle
3.3 Model Performance
3.4 Statistical Tests
4 Conclusions
References
Spatiotemporal Analysis of Agricultural Drought in Krishna River Basin
1 Introduction
2 Study Area
3 Materials and Methods
3.1 Data
3.2 Methodology
4 Results and Discussion
4.1 Temporal Analysis of SWDI
4.2 Aridity in KRB
4.3 Spatial Analysis of SWDI
5 Conclusions
References
Towards Imaging-based Quantification of Deterioration Using Colour Space Study
1 Introduction
2 HSV Colour Space
3 Materials
3.1 Cement
3.2 Fine Aggregate
3.3 Coarse Aggregate
3.4 Water
4 Experimental Programme
5 Results and Discussion
6 Conclusions
References
Trend Analysis of Climate Variables and Extremes Over Maner River Basin, India
1 Introduction
2 Materials and Methods
2.1 Study Area
2.2 Data Used
2.3 Methodology
3 Results
3.1 Observed Trends in the Seasonal and Annual Precipitation and Temperature
3.2 Observed Trends in the Climate Extreme Indices
4 Conclusions
References
Urban Dynamics and Impact Assessment of Bengaluru–Mysuru Expressway Corridor
1 Introduction
1.1 Introduction/Overview
1.2 Literature Review
2 Materials and Methods
2.1 Study Area
2.2 Data
2.3 Methodology
3 Results and Discussions
3.1 Land Use Classification
3.2 Future Land Use Transition
3.3 Sub-regional Analysis
3.4 Conclusion
References

Citation preview

Lecture Notes in Civil Engineering

Shashi Mesapam Anurag Ohri Venkataramana Sridhar Nitin Kumar Tripathi   Editors

Developments and Applications of Geomatics Proceedings of DEVA 2022

Lecture Notes in Civil Engineering Volume 450

Series Editors Marco di Prisco, Politecnico di Milano, Milano, Italy Sheng-Hong Chen, School of Water Resources and Hydropower Engineering, Wuhan University, Wuhan, China Ioannis Vayas, Institute of Steel Structures, National Technical University of Athens, Athens, Greece Sanjay Kumar Shukla, School of Engineering, Edith Cowan University, Joondalup, WA, Australia Anuj Sharma, Iowa State University, Ames, IA, USA Nagesh Kumar, Department of Civil Engineering, Indian Institute of Science Bangalore, Bengaluru, Karnataka, India Chien Ming Wang, School of Civil Engineering, The University of Queensland, Brisbane, QLD, Australia Zhen-Dong Cui, China University of Mining and Technology, Xuzhou, China

Lecture Notes in Civil Engineering (LNCE) publishes the latest developments in Civil Engineering—quickly, informally and in top quality. Though original research reported in proceedings and post-proceedings represents the core of LNCE, edited volumes of exceptionally high quality and interest may also be considered for publication. Volumes published in LNCE embrace all aspects and subfields of, as well as new challenges in, Civil Engineering. Topics in the series include: • • • • • • • • • • • • • • •

Construction and Structural Mechanics Building Materials Concrete, Steel and Timber Structures Geotechnical Engineering Earthquake Engineering Coastal Engineering Ocean and Offshore Engineering; Ships and Floating Structures Hydraulics, Hydrology and Water Resources Engineering Environmental Engineering and Sustainability Structural Health and Monitoring Surveying and Geographical Information Systems Indoor Environments Transportation and Traffic Risk Analysis Safety and Security

To submit a proposal or request further information, please contact the appropriate Springer Editor: – Pierpaolo Riva at [email protected] (Europe and Americas); – Swati Meherishi at [email protected] (Asia—except China, Australia, and New Zealand); – Wayne Hu at [email protected] (China). All books in the series now indexed by Scopus and EI Compendex database!

Shashi Mesapam · Anurag Ohri · Venkataramana Sridhar · Nitin Kumar Tripathi Editors

Developments and Applications of Geomatics Proceedings of DEVA 2022

Editors Shashi Mesapam Department of Civil Engineering National Institute of Technology Warangal Hanamkonda, Telangana, India Venkataramana Sridhar Department of Biological Systems Engineering Virginia Tech Blacksburg, VA, USA

Anurag Ohri Department of Civil Engineering Indian Institute of Technology (Banaras Hindu University) Varanasi, India Nitin Kumar Tripathi School of Engineering and Technology Asian Institute of Technology Bangkok, Thailand

ISSN 2366-2557 ISSN 2366-2565 (electronic) Lecture Notes in Civil Engineering ISBN 978-981-99-8567-8 ISBN 978-981-99-8568-5 (eBook) https://doi.org/10.1007/978-981-99-8568-5 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore Paper in this product is recyclable.

Foreword

It gives me immense pleasure to write this foreword to the “Lecture notes in Civil Engineering” containing the selected papers presented during the International Virtual Conference on Developments and Applications of Geomatics (DEVA—2022) held during 29th–31st August, 2022. The conference was held to coincide with the superannuation of Prof. Deva Pratap. NITW has been running an M.Tech. course in Remote Sensing and GIS, in the Department of Civil Engineering, since 2000 and has established itself as a pioneering institute in the field with its alumni occupying key positions. Many Ph.D. level research work and sponsored R&D works are in progress. I am happy to note that 38 peer-reviewed papers are part of this volume covering various topics where Remote Sensing and GIS can be applied. The National and International Advisory Committee consisted of experts who have been recognised for their contribution. I am sure the wealth of information contained in this volume will be useful for the Geomatics engineers who use a wide range of technologies for application in various fields like Civil Engineering, Computer Engineering and Software engineering covering a wide range of topics like land development and planning, satellite and information technologies, advanced surveying etc. The topics in Geoinformatics are interdisciplinary in nature. I am confident that this volume will be very useful and add to the knowledge in the interdisciplinary area of Geoinformatics. I convey my appreciation to the Organizing Secretaries of the Conference for the excellent work done. Warangal, India

Prof. Bidyadhar Subudhi Director

v

Preface

National Institute of Technology Warangal, the first among the chain of erstwhile Regional Engineering Colleges has a well established Department of Civil Engineering which runs an undergraduate and 8 Master degree programs besides Ph.D. One of the specializations is the Remote Sensing and GIS, started in the year 2000. Over the years, this programme has grown in stature and attracts many research projects. A conference was organised to coincide with the superannuation of Prof. Deva Pratap, who initiated the starting of the Master degree program at NIT Warangal, and Shashi Mesapam, Keesara Venkatareddy and Manali Pal being the Organizing Secretaries. The conference attracted researchers from all parts of India as well as other countries. The conference had to be conducted in a virtual mode due to Covid. The papers covered various areas of application of Geomatics like water resources, environmental engineering, ecosystem management, structural health monitoring, transportation engineering, web GIS etc. The selected papers presented in the conference form the contents of this Lecture Note Series. We are sure the contents of this volume will be useful for researchers working in the area of Remote Sensing and GIS. Hanamkonda, India Varanasi, India Blacksburg, USA Bangkok, Thailand

Shashi Mesapam Anurag Ohri Venkataramana Sridhar Nitin Kumar Tripathi

vii

Acknowledgements

A conference of this nature covering a wide area cannot be organized without the involvement of many individuals and organizations. We thank the Director and administration of NIT Warangal for all the infrastructure support. The conference had excellent national and international advisory committee for critically reviewing the papers and offering constructive suggestions for improving the quality of papers selected for the conference. Special mention must be made of IITT Navishkar I-Hub Foundation (IITTNiF) and GIS Monk whose financial support made the conference very successful. Organizing Secretaries Shashi Mesapam Keesara Venkatareddy Manali Pal

ix

Contents

Analyzing the Potential Application of Low-Cost Digital Image Correlation in Direct Shear Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Alhakim, C. Nuñez-Temes, J. Ortiz-Sanz, and M. Arza-García

1

Applications of GIS in Estimating the Probable Maximum Earthquake Magnitude for Amaravati Region, Andhra Pradesh, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. Madhusudhan Reddy, R. Siddhardha, G. Kalyan Kumar, and R. Suresh

15

Assessing the Effect of Land Use Land Cover Change on the Water Quality Index of a River Basin Using GIS and Remote Sensing Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . W. S. Adhima, J. S. Gouri, Pooja N. Raj, P. S. Riya, and Lini R. Chandran

25

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground Water Levels in Kurukshetra, Haryana . . . . . . . . . . . . . . . . . . . . . . Vikas Singh and A. K. Prabhakar

43

Assessment of Land Use—Land Cover Changes in District Dehradun (1991–2021) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Madhusudan Thapliyal and A. K. Prabhakar

55

Comparison of Streamflow Simulations for Different DEMs . . . . . . . . . . . Nagireddy Venkata Jayasimha Reddy and R. Arunkumar Comprehensive Analysis of Impact of COVID-19 Lockdown on Air Quality in Andhra Pradesh, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . Donthi Rama Bhupal Reddy and Ramannagari Bhavani Development of Mobile Application for Assessing Urban Heat Island (UHI) Using Geospatial Techniques a Case Study of Chennai City . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. Jayalakshmi

69

79

95

xi

xii

Contents

Drones as an Alternate Communication System During Calamities . . . . . 109 D. S. Vohra, Pradeep Kumar Garg, and Sanjay Kumar Ghosh Drought Analysis of an Area Using Google Earth Engine . . . . . . . . . . . . . . 123 Jyothsna Devi Adapa and Keesara Venkatareddy Effects of Urbanization on Land Use Land Cover of Warangal Region Using RS and GIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Ch. Sree Laxmi Pavani, Keesara Venkatareddy, and S. Joshmitha Effect of LULC Changes on Land Surface Temperature . . . . . . . . . . . . . . . 155 Rajashekar Kummari, Pavan Kumar Reddy Allu, Shashi Mesapam, Ayyappa Reddy Allu, Bhargavi Vinakallu, and Bhanu Prakash Ankam Estimation of Aerosol Direct Radiative Forcing in Southern India . . . . . . 175 K. Tharani, Deva Pratap, Keesara Venkatareddy, and P. Teja Abhilash Estimation of Groundwater Potential Zones in Southern Dry Agro-Climatic Area Using Geoinformatics and AHP Technique . . . . . . . . 185 A. B. Gireesh and M. C. Chandan Evaluation and Prediction of Land Use and Land Cover Changes in the Kumaradhara Basin, Western Ghats, India . . . . . . . . . . . . . . . . . . . . 201 N. Roopa, N. Namratha, H. Ramesh, and K. C. Manjunath Evaluation of Surface Soil Moisture Using Remote Sensing and Field Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 T. N. Santhosh Kumar and Abhishek A. Pathak Evaluation of the Influence of Land Use and Climate Changes in Runoff Simulation Using Semi-Distributed Hydrological Model . . . . . 231 M. S. Saranya and Vinish V. Nair Flood Damage Assessment of a River Basin Using HEC-GeoRAS . . . . . . 245 K. C. Amal Vishnu and Vinish V. Nair Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 Sampath Kumar, Talari Reshma, Savitha Chirasmayee, Kasa Priyanka, Kokku Priyanka, and Gokla Ram GIS and RS-Based Soil Erosion and Sediment Yield Modelling in Manikpur, Chhattisgarh, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 B. Himajwala and A. D. Prasad Groundwater Level Trends Over Southern India . . . . . . . . . . . . . . . . . . . . . 289 Kotapati Narayana Loukika, Keesara Venkatareddy, and Eswar Sai Buri Impact of Climate Change on Streamflow Over Nagavali Basin, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 Nageswara Reddy Nagireddy and Keesara Venkatareddy

Contents

xiii

Impervious Surface Area Prediction Using Landsat Satellite Imagery and Open Source GIS Plugin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Ayyappa Reddy Allu and Shashi Mesapam Influence on Water Characteristics Away from Various Sources of NIT Kurukshetra Using ArcGIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Rahul Deopa and K. K. Singh Landslide Hazard Zonation Mapping Using Remote Sensing and GIS in Mountainous Terrain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 Dolonchapa Prabhakar, Anoop Kumar Shukla, Babar Javed, and Satyavati Shukla Modeling Daily Streamflow from Idamalayar Catchment Using SWAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 C. Reshma and R. Arunkumar Modelling the Low Impact Development Alternatives for Rainfall Runoff Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 B. Aneesha Satya, M. Shashi, and Allu Pavan Kumar Reddy Performance Evaluation of Support Vector Machine and Random Forest Techniques for Land Use-Land Cover Classification—A Case Study on a Mili Scale Agricultural Watershed, Tadepalligudem, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379 Chirasmayee Savitha and Talari Reshma Photogrammetric Survey of an Intertidal Area: A Case Study in NW Spain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393 M. Gil-Docampo, S. Peña-Villasenín, S. Peraleda-Vázquez, R. Carballo, and N. Gómez-Conde Potential Zones Identification to Effectively Exploit Solar and Wind Energy in the State of Assam—A MCDA Approach Using GIS and Remote Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 P Taniya Raj and N. S. R. Prasad Prediction of Soil Organic Carbon in Unscientific Coal Mining Area Using Landsat Auxiliary Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427 Naorem Janaki Singh, Lala I. P. Ray, Sanjay-Swami, and A. K. Singh Rainfall Runoff Modeling Using HEC-HMS for Munneru River Basin, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441 Eswar Sai Buri, Keesara Venkatareddy, and K. N. Loukika Spatio-Temporal Surface Urban Heat Island Effect Analysis Over Tiruchirappalli City, India, Using GIS Techniques . . . . . . . . . . . . . . . . . . . . 449 K. S. Arunab, Ajay Badugu, Aneesh Mathew, and Padala Raja Shekar

xiv

Contents

Simulation of Streamflow and the Assessment of Nutrient Loadings for the Indravati River Basin of India using SWAT . . . . . . . . . . . . . . . . . . . 467 Ch. Venkateswarlu, R. Manjula, P. Yuvaraja, and S. Hemavathi Spatiotemporal Analysis of Agricultural Drought in Krishna River Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485 Hussain Palagiri and Manali Pal Towards Imaging-based Quantification of Deterioration Using Colour Space Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 V. Guru Prathap Reddy, K. Bhanu, T. Tadepalli, and Rathish Kumar Pancharathi Trend Analysis of Climate Variables and Extremes Over Maner River Basin, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509 Koppuravuri Ramabrahmam and Keesara Venkatareddy Urban Dynamics and Impact Assessment of Bengaluru–Mysuru Expressway Corridor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519 S. Suhas, V. Bhavani, B. M. Vishwanath, Ruthvik Krishna, and M. C. Chandan

About the Editors

Prof. Shashi Mesapam is currently an Associate Professor at the Department of Civil Engineering, National Institute of Technology Warangal, India. He obtained his B.Tech. (Civil) from Osmania University, Hyderabad, and M.Tech. (Remote Sensing and Photogrammetric Engineering) and Ph.D. from the Indian Institute of Technology, Roorkee. His major areas of research interests include Photogrammetry, Unmanned Aerial Vehicles (UAV) and its Applications, Digital Image Processing, Remote Sensing, Advanced Surveying. He has published 26 papers in reputed international journals and 4 book chapters in reputed publications. Prof. Anurag Ohri is currently an Associate Professor at the Department of Civil Engineering, Indian Institute of Technology (Banaras Hindu University), Varanasi, India. He obtained his B.Tech. (Civil) from Regional Engineering College, Kurukshetra, and M.Tech. (Remote Sensing and Photogrammetric Engineering) from the Indian Institute of Technology, Roorkee and Ph.D. from the Indian Institute of Technology (Banaras Hindu University). His major areas of research interests include Mapping, Development of GIS based Decision support systems, GIS application in Smart Cities, Land use planning. He has published 34 papers in reputed international journals and 3 book chapters in reputed publications. Prof. Venkataramana Sridhar is currently an Associate Professor at the Biological Systems Engineering in Virginia Tech, Blacksburg, USA. He obtained his B.E. in Agricultural Engineering from Tamil Nadu Agricultural University, India and M.Eng. in Irrigation Engineering & Management from Asian Institute of Technology, Thailand and Ph.D. from Oklahoma State University, Stillwater, Oklahoma, USA. His major areas of research interests include Global change; urban systems and sustainability; agricultural water management; water resources systems; modeling; extreme events (floods and droughts); food-energy-water nexus; developing, implementing and evaluating earth system models; new hydrologic instrumentation, sensors, drones, databases, and model applications for use by federal, state, and private stakeholders for real-time operational applications; international

xv

xvi

About the Editors

activities. He has published 90 papers in reputed international journals and 6 book chapters in reputed publications. Prof. Nitin Kumar Tripathi is currently a Professor at the Asian Institute of Technology, Thailand. He obtained his B.Tech. (Civil) from National Institute of Technology, Warangal, India, and M.Tech. (Remote Sensing) and Ph.D. from the Indian Institute of Technology, Kanpur. He has published 211 papers in reputed international journals, two books and 12 book chapters in reputed publications. He is currently Editor-in-chief of International Journal of Geoinformatics and Editorial Board Member of Int. Jour. of Applied earth Observation and Geoinformation (ITC Journal), Asian J. of Geoinformatics, Indian J. of Remote Sensing and Photogrammetry, Journal of Prince of Songkhla University, West Indian Journal of Civil Engineering, International Journal of Geoinformatics, Arabian Journal for Science and Engineering, International Journal of Remote Sensing—Taylor and Francis, International Journal of Imaging, J. of Chemistry and Environment, Geojournal— Springer, Disaster Advances, Natural Hazards, International Journal of Geographical Information Science.

Abbreviations

σ % °C μ 2D 3D 95PPU ac ADRF AHP AI ALPHA_BF ALPHA_BNK AMD AMJ ANN AP API APK AQI ASCAT ASMR ASTER AT ATC AWBGT AWS B BBMP BMIC

Standard Deviation Percent Degree Celsius Micro Two Dimensional Tridimensional 95 Percent Prediction Uncertainty Acres Aerosol Direct Radiative Forcing Analytic Hierarchy Process Aridity Index Base Flow Alpha Factor Base Flow Alpha Factor For Bank Storage Acid Mine Drainage April–June Artificial Neural Network Andhra Pradesh Application Programming Interface Android Package Kit Air Quality Index Advanced Scatterometer Advanced Microwave Scanning Radiometer-2 Advanced Space Borne Thermal Emission and Reflection Radiometer Ambient Temperature Air Traffic Control Approximated Wet Bulb Globe Temperature Annual Maximum Series Blue Bruhat Bengaluru Mahanagara Palike Bengaluru Mysuru Infrastructure Corridor xvii

xviii

BNU-ESM BOD BSI BTA BW C CA Ca CaCO_3 CBD CCME CDD CGWB CH_K2 CH_N2 CI CIE Cl Cm CMI CMIP CMOS CN CNRM-CM5 CO Conc COVID CP CPCB CR CRS CRUTS CWC CWD CWT D dB DEM DIC DN DO

Abbreviations

Beijing Normal University Earth System Model Biochemical Oxygen Demand Bare Soil Index Back Trajectory Analysis Bandwidth Crop Management Factor Cellular Automata Calcium Calcium Carbonate Central Business District Canadian Council Of Ministers of the Environment Maximum Number of Consecutive Days When Daily Rainfall < 1 Mm Central Ground Water Board Effective Hydraulic Conductivity in Main Channel Alluvium Manning’s N Value for the Main Channel Consistency Index Commission for International Illumination Chlorides Centimeters Crop Moisture Index Coupled Model Inter-Comparison Project Complementary Metal-Oxide-Semiconductor Curve Number Centre National for Meteorological Research Coupled Model Version 5 Carbon Monoxide Concentration Corona Virus Disease Checkpoints Central Pollution Control Board Consistency Ratio Coordinate Reference System Climate Research Unit Central Water Commission Maximum Number of Consecutive Days When Rainfall > 1 Mm Concentration Weighted Trajectory Distance from Drainage Network Decibel Digital Elevation Model Digital Image Correlation Digital Number Dissolved Oxygen

Abbreviations

DTM EC EDRIR EPCO EPSG ERAinterim ERSCAT ESA ESCO ESDAC ESDS ESRI ESSMI ET ETM+ EUFD F FAO FC FCC FDM Fig FIM FLOWA FR G G ratio GCC GCM GCP GDVI GHG GHI GIMMS GIS GNDVI GNSS GPCC GPM GPS GRD GRVI GS GSAVI GSD

xix

Digital Terrain Model Electrical Conductivity Earthquake Disaster Risk Index Report Plant Uptake Compensation Factor European Petroleum Survey Group European Reanalysis Interim European Remote Sensing Satellite Scatterometer European Space Agency Soil Evaporation Compensation Factor European Soil Data Center Earth Science Data System Environmental Systems Research Institute Empirical Standardized Soil Moisture Index Evapotranspiration Enhanced Thematic Mapper Plus European Union Floods Directive Flow Accumulation Food and Agriculture Organization Field Capacity False Colour Composite Fused Deposition Modeling Figure Flood Inundation Mapping Fuzzy Logic Ordered Weight Averaging Frequency Ratio Green Green Ratio Gnu Compiler Collection General Circulation Models Ground Control Point Green Difference Vegetation Index Green House Gas Global Horizontal Irradiance Global Inventory Modeling and Mapping Studies Geographic Information System Green Normalized Difference Vegetation Index Global Navigation Satellite System Global Precipitation Climatology Centre Ground Water Potential Mapping Global Positioning System Ground Range Detected Green Red Vegetation Index Ground Station Green Soil Adjusted Vegetation Index Ground Sample Distance

xx

GUI GW GW_DELAY GW_REVAP GWDRVI GWL GWPI GWPZ GWQMN H H+ HCO_3 HD HEC HEC-GeoRAS HEC-RAS HI HMS HR HRU HWSD HYSPLIT IIASA IMD IPCC IPSL-CM5A-MR IRDB ISC ISR ISRIC IT IUCN JAS JFM Kc Khist kHz KIADB Kloc KM KrRB KaRB

Abbreviations

Graphical User Interface Ground Water Groundwater Delay Groundwater Revap Coefficient Green Wide Dynamic Range Vegetation Index Ground Water Level Ground Water Potential Index Ground Water Potential Zone Threshold Depth of Water in the Shallow Aquifer Required For Return Flow (Mm) Hue Hydrogen Ion Bicarbonate High Definition Hydrologic Engineering Centre Hydrologic Engineering Center’s Geospatial River Analysis System Hydrologic Engineering Center’s River Analysis System Horizontal Irradiance Hydrologic Modelling System Hazard Ranking Hydrologic Response Unit Harmonized World Soil Database Hybrid Single Particle Lagrangian Integrated Trajectory International Institute for Applied Systems Analysis Indian Meteorological Department Intergovernmental Panel on Climate Change Institute Pierre-Simon Laplace Coupled Model Version 5A-Medium Resolution Irrigation Design and Research Board Indian Standard Code Impervious Surface Ratio International Soil Reference and Information Centre Information Technology International Union for Conservation of Nature July–August January–March Kappa Coefficient Kappa Histogram Kilohertz Karnataka Industrial Areas Development Board Kappa Local Kilometers Krishna River Basin Karamana River Basin

Abbreviations

KSB KSPCB L Lat LAT Q LC Long LOS LR LS LST LULC LWGNT LWGNTCLN LWTUP LWTUPCLN M M(max^obs ) Max MC Mc MCDA MCDM Mg Mg/L MGSAVI MHz Min MK ml MLC Mmax MOLUSCE MSAVI MT MtP mW Mw Na NASA NBBSLUP NCR NDBI NDMA

xxi

Kijko-Sellevoll-Bayes Kerala State Pollution Control Board Length Latitude Lateral Flow Land Cover Longitude Visible Line Of Sight Logistic Regression Slope Length and Slope Steepness Factor Land Surface Temperature Land Use Land Cover Surface Net Downward Longwave Flux Surface Net Downward Longwave Flux Assuming No Aerosol Upwelling Longwave Flux at TOA Upwelling Longwave Flux at TOA Assuming No Aerosol Meters Observed Maximum Magnitude Maximum Markov Chain Magnitude Completeness Multiple-Criteria Decision Analysis Multi Criteria Decision Making Magnesium Milligram per Liter Modified Green Soil Adjusted Vegetation Index Megahertz Minimum Mann & Kendall Microliters Maximum Likelihood Classification Probable Maximum Earthquake Magnitudes Modules for Land Use Change Evaluation Second Modified Soil Adjusted Vegetation Index Metric Tons Meteorological Parameters Milliwatt Moment Magnitude Sodium National Aeronautics and Space Administration National Bureau of Soil Survey and Land Use Planning Net Change Ratio Normalized Difference Built-Up Index National Disaster Management Authority

xxii

NDT NDVI NDWI NEX-GDDP NGT NHAI NIR NLOS NO2 NPCM NPI NRSC NSE NSF OA OBD OLI OND OSM P PA PBIAS PCM PDSI PET PLA Plat PM PMP Ppm PRCPTOT PSCF PSO PtClx 4 PTFs PTL PV PVIP QGIS QGIS qp Qp QSWAT

Abbreviations

Non Destructive Testing Normalized Difference Vegetation Index Normalized Difference Water Index NASA Earth Exchange Global Daily Downscaled Climate Projections National Green Tribunal National Highway Authority of India Near Infra-Red Non Visible or Beyond Line Of Sight Nitrogen Dioxide Normalized Pairwise Comparison Matrix National Pollution Index National Remote Sensing Centre Nash-Sutcliffe Simulation Efficiency National Sanitation Foundation Overall Accuracy Over Burden Dump Operational Land Imager October–December Open Street Map Support Practice Factor Producers Accuracy Percentage Bias Pairwise Comparison Matrix Palmer Drought Severity Index Potential Evapotranspiration Polylactic Acid Platinum Particulate Matter Probable Maximum Precipitation Parts Per Million Total Amount of Rainfall in Wet Days Potential Source Contribution Factor Particle Swarm Optimisation Chloroplatinate Ion Pedotransfer Functions Power Transmission Lines Photovoltaic Periyar Valley Irrigation Project Quantum Geographic Information System Quantum GIS Peak Discharge Per Sq. Km Area of Catchment Peak Discharge of the Unit Hydrograph for the Catchment Area QGIS Interface for Soil and Water Assessment Tool

Abbreviations

r R Rd R10MM R2 R20MM R95PTOT RCP RECHARGE_DP REVAPMN RF RGB RH RI RI RMSE ROI RRMSE RS RSR RTK RUSLE RX1DAY RX5DAY S S S/cm SO4 SAR SAVI SCS SDG SDR SED YIELD SEZ SfM SLEUTH SLR SM SMA SMADI SMAP

xxiii

Correlation Coefficient Rainfall Erosivity Factor Red Number of Days When RR > 10 mm. Count the Number of Days Where Rrij > 10 mm Coefficient of Determination Number of Days When RR > 20 mm. Count the Number of Days Where Rrij > 20 mm Total Rainfall When RR > 95p Representative Concentration Pathway Deep Aquifer Percolation Factor Threshold Depth of Water for Revap or Percolation to Occur Random Forest Red Green Blue Relative Humidity Rainfall Intensity Redness Index Root Mean Square Error Region of Interest Relative Root Mean Square Error Remote Sensing Root Mean Square Error to Standard Deviation Ratio Real Time Kinematic Revised Universal Soil Loss Equation Monthly Maximum 1-Day Precipitation Monthly Maximum Consecutive 5-Day Precipitation Slope Saturation Simen Per Centimeter Sulphate Synthetic Aperture Radar Soil Adjusted Vegetation Index Soil Conservation Service Sustainable Development Goal Sediment Delivery Ratio Sediment Yield Special Economic Zone Structure from Motion Slope Land Use Excluded Urban Transport Hillshade Stepwise Linear Regression Soil Moisture Soil Moisture Accounting Soil Moisture Agricultural Drought Index Soil Moisture Active Passive

xxiv

SMDI SMI SMOS SMU SNAP SNR SO2 SOC SOL_AWC SOL_BD SOL_BD SOL_K SPEI SPI SPSS SR1 SR2 SR3 & SR4 SSA SSM SSR STD STL SUFI 2 SUH SUR Q SURLAG SWAT SWAT CUP SWDI SWGNT SWGNTCLN SWIR SWMM SWTNT SWTNTCLN SYobd SYpred T TA TB TC TCU TDS TH

Abbreviations

Soil Moisture Deficit Index Soil Moisture Index Soil Moisture and Ocean Salinity Soil Mapping Unit Sentinel Application Platform Signal to Noise Ratio Sulphur Dioxide Soil Organic Carbon Available Water Capacity of the Soil Layer Moist Bulk Density Mm Layer Moist Bulk Density (G/Cm3) Saturated Hydraulic Conductivity Standardized Precipitation Evapotranspiration Index Standardized Precipitation Index Statistical Package for the Social Sciences Scenario 1 2 3 & 4 Seismic Study Area Surface Soil Moisture Surface Soil Roughness Standard Deviation Social and Travel Lockdown Sequential Uncertainty Fitting 2 Synthetic Unit Hydrograph Surface Runoff Surface Runoff Lag Coefficient Soil and Water Assessment Tool SWAT Calibration and Uncertainty Programs Soil Water Deficit Index Surface Net Downward Shortwave Flux Surface Net Downward Shortwave Flux Assuming No Aerosol Short Wave Infra Red Storm Water Management Model Toa Net Downward Shortwave Flux TOA Net Downward Shortwave Flux Assuming No Aerosol Observed Sediment Yield Simulated Sediment Yield Temperature Total Alkalinity Time Base of SUH Total Coliform True Colour Unit Total Dissolved Solids Total Hardness

Abbreviations

TIN TIRS TM Tm TNN TNX TOA tp TSA TWI TXN TXX UA UAS UAV UER UGM UHI UR USDA USGS UTM V VCI VES VIC VSWI W50 W75 WAD WAWQI WBGT WDVI Wf WG WGS WHO WMO WOA WOE WP WQ WQI

xxv

Triangular Irregular Network Thermal Infrared Sensor Thematic Mapper Peak Time of SUH Monthly Minimum Value of Daily Minimum Temperature Monthly Maximum Value of Daily Minimum Temperature Top of Atmosphere Basin Lag Theil-Sen Approach Topographic Wetness Index Monthly Minimum Value of Daily Maximum Temperature Monthly Maximum Value of Daily Maximum Temperature Users Accuracy Unmanned Aerial System Unmanned Aerial Vehicle Urban Expansion Rate Urban Growth Model Urban Heat Island Urbanization Rate United States Department of Agriculture United States Geological Survey Universal Transverse Mercator Value Vegetation Condition Index Vertical Electrical Sounding Variable Infiltration Capacity Vegetation Supply Water Index Width of SUH at 50% of Peak Discharge Ordinate Qp in Hours Width of SUH at 75% of Peak Discharge Ordinate Qp in Hours World Atlas of Desertification Weighted Arithmetic Wqi Wet Bulb Globe Temperature Weighted Difference Vegetation Index Weightage Factor Western Ghats World Geodetic System World Health Organization World Meteorological Organization Weighted Overlay Analysis Weight of Evidence Wilting Point Water Quality Water Quality Index

xxvi

WR50 WR75 WRIS WRS WS Xm XML Xn Xn-m XS YLCD Z_MK

Abbreviations

Half Width of SUH at 50% of Peak Discharge Ordinate Qp in Hours Half Width of SUH at 75% of Peak Discharge Ordinate Qp in Hours Water Resource Information System Worldwide Reference System Wind Speed Highest Value in the Series Extensible Markup Language Mean of N Annual Maximum Values Mean of the Series Excluding the Highest Value Cross Section Yearly Land Cover Dynamic Standardized MK-Test Statistic

Analyzing the Potential Application of Low-Cost Digital Image Correlation in Direct Shear Test G. Alhakim, C. Nuñez-Temes, J. Ortiz-Sanz, and M. Arza-García

Abstract Nowadays, Digital Image Correlation (DIC) has become a frequently used optical technique for testing lab materials due to its nature of non-contact method, allowing for full-field displacements and strains to be accurately measured. The purpose of this study is to examine the potential employment of DIC in performing the standard direct shear test by using a modified box to observe and measure the shear properties of a soil sample. Accessible tools (e.g., a consumer grade DSLR camera) and DIC open-source software were employed in this study in order to monitor every portion of the speckled pattern during deformation and tracking the relative displacements. While it is often difficult to separate the relative contributions of individual error sources in any optical technique, we propose the use of two different validation methods for assessing the accuracy of DIC: the noise-floor and the direct comparison with a second trusted source, like transducers. The results of the conducted study demonstrate that the DIC technique implemented on a properly prepared direct shear laboratory setup opens up new possibilities for the effective and accurate analysis of deformations in soil materials under direct shear loading conditions. The calculations of the noise floor of the setup and speckled pattern in terms of the mean and distribution of the displacements showed validated results with STD of 0.001 and 0.0012 mm for the horizontal and vertical planar components, respectively. The displacements measured by DIC showed good agreement with the results of the transducer with an average error of 0.1 mm. Keywords Direct shear test · DIC analysis · Displacement measurement · NDT

G. Alhakim (B) · C. Nuñez-Temes · J. Ortiz-Sanz · M. Arza-García Civil and Geomatics Research Group, Department of Agroforestry Engineering, University of Santiago de Compostela, Santiago, Spain e-mail: [email protected] G. Alhakim Civil and Environmental Engineering Department, Faculty of Engineering, Beirut Arab University, Beirut, Lebanon © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_1

1

2

G. Alhakim et al.

1 Introduction In recent years, Digital image correlation (DIC) has become one of the preferred optical techniques for the analysis of close-range displacements in several lab and in-field experiments, being able to capture the shape, motion, and deformation of object surfaces. Since the development of these DIC methods over the past decades, numerous applications have been found in many fields and areas of research, due to its ease of use. Hence, this technique has found a particularly interesting application niche in those fields related to civil engineering. The basic principle of DIC is to track a speckle pattern in a sequence of images and to calculate the displacements of small portions of this pattern (often called subsets) between one reference image and the other (deformed) images. The variations of coordinates between the original and deformed subset centers provide the required displacement vector of the targeted point. Similarly, the full-field deformation of other interrogated points could be acquired. For this purpose, DIC usually requires an artificial speckle pattern to resolve unique issues [1]. Nonetheless, the pattern can also sometimes be naturally present over the surface of some materials (e.g., rocks, sand, etc.). In general, there are now multiple software solutions, both commercial and open source, to perform the correlation between the images. Among others, we can find commercial systems like Aramis (GOM) or VIC-3D (Correlated Solutions), that are normally provided with their own specialized hardware and optical components. As the use of DIC becomes more and more widespread, diverse open-source projects are also emerging (e.g., nCorr [2], py2DIC [3], etc.) making this technique more accessible, even for small companies and laboratories with more limited resources. At a level of hardware, also industrial high-speed cameras have proven to be replaceable—at low cost—by consumer grade cameras in certain experimental setups where the frame rate is not a decisive issue [4]. The affordability of the approach and the simplicity of its implementation have democratized to some extent the use of DIC, bringing it closer to all practitioners. DIC is currently considered a simple but reliable technique for analyzing surface displacements and deformations. However, the analysis of the results and their accurate assessment can be technically challenging. The main problem is a general lack of control over the final quality of the output, as the users are sometimes more fascinated by the detail of the full-field maps rather than aware of the metric quality of the results obtained. It is important to remark that DIC measurements, as many other optical techniques, are affected by different types of errors. Among others, the results can be affected by the stationary noise from camera sensors, errors related to the illumination or temperature, errors associated with speckle pattern, errors associated with the processing, external biases (e.g., vibrations), etc. [5, 6]. Therefore, the accuracy of every experiment requires careful verification.

Analyzing the Potential Application of Low-Cost Digital Image …

3

1.1 DIC Applications in Lab Testing of Materials The experimental technique of imaging correlation has gained widespread acceptance as a reliable tool for strain measurements. Thus, DIC has been appealing to researchers in various engineering applications as a non-destructive testing (NDT). Many studies have examined the use of the DIC technique to analyze the concrete behavior, such as the distribution of deformation and crack propagation of concrete [7], quantification of the fracture properties [8] and fatigue behavior [9] of reinforced concrete beams. Besides, the DIC optical method was employed to assess the behavior of road materials [10–12]. Furthermore, geotechnical engineering investigation relies greatly on DIC practices to determine subpixel displacements in the soil, through which strain is estimated [13]. Plé, Tourabi and Abuaisha [14] have investigated the strain deformation during a direct tensile test performed on a clayey soil by applying 3-D DIC technique. They stated that this method could be adapted for the strain determination of the cap cover in landfills composed of compacted clay. A modified consolidation mold has been developed to monitor the micro-structure behavior of unsaturated soil with axial drainage conditions. By using this half mold oedometer apparatus, Liu et al. [15] studied the deformation field of silty clay soil by applying digital image analysis. They declared the importance of this study in terms of understanding the correlation between micro and macro mechanical properties of the soil. Ko et al. [16] conducted an experiment of a small-scale retaining wall by reproducing the collapse behavior of the structure and measured the deformation and displacement of the whole surface of the wall by using 3D-DIC. Tong and Yoo [17] discussed the application of DIC technology on geotechnical small-scale models such as a retaining wall, trapdoor, shallow footing, and tensile test on geogrid. They revealed the effectiveness of DIC in monitoring the deformation and strain field during these laboratory tests.

1.2 Direct Shear Test The direct shear test is one of the oldest experimental procedures implemented in geotechnical engineering practice, and it is the most widely used geotechnical shear device due to its simplicity. This experimental procedure is conducted in order to determine the shear strength and shear parameters of soil materials. The accurate determination of these parameters represents a key issue in the design of various structures, e.g., retaining walls, earth buildings, dams, and roads, as well as the design of foundations beneath any architectural and/or engineering structure. However, one of the demerits of the direct shear test could be the non-uniform distributions of stress and strain on the failure plane. Accordingly, numerical methods have been employed to analyze the details of stress and strain occurrence within a shear box, that would be difficult to measure in

4

G. Alhakim et al.

the experimental study. Dirgelien˙e, Skuodis and Griguseviˇcius [18] carried out experimental and numerical investigations of the direct shear test performed under constant vertical stress and constant sample volume. The results of the finite element method showed that the vertical stress and displacements were non-uniformly distributed in the sample in both conditions. Zhang and Thornton [19] simulated the direct shear test by using the discrete element method and illustrated the heterogeneous distributions of stress and strain. In addition, they demonstrated that the shear strain as well as the dilation concentrated in the mid-height of the specimens (within the shear zone). To study the soil behavior at the microscopic level, Yan [20] presented a 3-D numerical model of granular soil with elongated particles in a direct shear test. It was observed that a localized dilative shear zone was revealed along the failure plane and the change in particle orientation was clearly noticed. While these non-uniform distributions of stress and strain can be numerically modelled, the employment of a full-field measurement NDT method like DIC could be very useful for tracking experimentally the changes in displacements on the lateral surface of the sheared samples and for validating the models [21]. The main limitation for applying DIC in the direct shear test is that the mold used in the test apparatus is a metal closed box, which makes it impossible to monitor the real behavior of the soil. Kong, Cheng, and Hua [22] applied digital image technique to direct shear test by employing a modified shear box and determined the displacement and strain fields as well as the particle orientation. They concluded that the maximum shear strain occurred at the interface between the upper and lower shear boxes and the formation of shear band could be clearly detected. However, in their research they did not consider any method to validate the accuracy of DIC. The review of the mentioned studies enlightens the benefits of DIC application in the direct shear test to monitor the strain field and displacement instead of assuming uniform distribution of stress and strain during shearing. Despite this, the usage of the DIC technique in the direct shear test of soil has not been fully investigated yet, especially when it comes to quantify the precision of this technology in this experimental test. Hence, the main objective of this research is to study the potential application of DIC in conducting the direct shear test on soil. Accordingly, a modified shear box and mold were designed for this purpose and accessible tools (e.g., a consumer grade digital single-lens reflex (DSLR) camera and open-source software) were employed. The relative contributions of individual error sources are hardly separated in any optical practice. Therefore, we suggest the use of two different validation methods to quantify the accuracy of DIC: the noise-floor, and the direct comparison with a conventional contact measurement device.

Analyzing the Potential Application of Low-Cost Digital Image …

5

Fig. 1 Shear mold and box in 3-D model

2 Materials and Method 2.1 Shear Mold In order to monitor the soil grains displacement and the deformation during shearing, the original shear box must be modified by cutting one of the sides of the upper and lower molds. For experimental purposes, and to prevent the damage of the original metal mold, a new cost-effective one was produced by applying Fused Deposition Modeling (FDM) printing. A 3-D printed mold was manufactured with a Prusa i3 (BQ, Spain) 3-D printer by using a bioplastic filament material, namely, Polylactic Acid (PLA). The original dimensions of the shear box (60 × 60 mm) were maintained. Moreover, the open sides of the upper and lower boxes were replaced with two plexiglass plates of 3 mm thickness, as illustrated in Fig. 1.

2.2 2D-DIC Fundamentals To be able to analyze displacements with 2D-DIC in the image series, the first stage consists in selecting a subset or portion of the speckled pattern to be tracked. The subset size should be chosen in a way that allows to distinguish every subset in the region of interest from all other subsets. There is no one-size-fits-all rule in this sense, but it is commonly accepted between DIC practitioners to choose a subset size that allows facets containing at least 3 speckles. Larger subsets typically result in lower displacement noise, but often at the cost of increased spatial smoothing [23]. A process of image correlation is performed to detect homologous subsets between the reference image and the deformed ones.

6

G. Alhakim et al.

As illustrated in Fig. 2, the coordinates of the central point of every initial subset can be mapped in the target subset by using the computed image displacement and first-order shape functions, representing translations, rotations, normal and shear strains (Eqs. 1 and 2). ς 1 = u + (∂u∂x)dx + (∂u∂v)dy

(1)

η1 = v + (∂v∂x)dx + (∂v∂y)dy

(2)

where ζ1 and η1 are the displacements of the subset; u and v represent the translations; ∂u/∂x and ∂v/∂y represent the normal strains; ∂u/∂y and ∂v/∂x represent the shear strains; and dx and dy represent the distances from the subset center to an arbitrary point within the same subset in the x and y directions, respectively. As DIC software works with pixel units, the displacement maps resulting from processing all the points of interest do not have an inherent scale. For this reason, DIC requires a calibration procedure to establish the image scale, relating pixel units with real units (what can be even done with a simple known distance). However, at the same time, the calibration procedure can be also used to correct for lens distortions if a calibration plate is used. In this study, we performed the 2D-DIC calibration by taking images of a special calibration plate (i.e., symmetric dot grid target) prior to the experiments. By tracking the displacements of the points of the grid in along the images of the plate, the parameters of the camera can be calibrated. If the intrinsic and extrinsic parameters of a 2D, single camera system are calibrated, then an outof-plane tilt of the test piece can be determined and corrected. In any case, when only one camera is used for DIC (2D-DIC) is crucial to ensure that the translation remain strictly perpendicular to the optical axis; otherwise, false displacements due to out-of-plane motion will be produced. Fig. 2 Schematic illustration of the basic principles in 2-D DIC

Analyzing the Potential Application of Low-Cost Digital Image …

7

2.3 Experimental Setup The first step to process a DIC experiment is to prepare a pattern on the surface of the analyzed sample. Therefore, to get a reference for the box displacements, a half of the transparent plate was painted with white paint and speckled randomly with black spots, as is typical for DIC, in order to maximize the contrast of the speckles and to improve the reliability of the tracked displacements. To monitor the experiments, image series were taken by using a digital single-lens reflex camera (DSLR) Canon EOS 1200D with a resolution of 18.1-megapixel, equipped with Canon EF lens (image settings: focal length 21 mm, ISO-100, exposure time 1/25s, f/6). The camera and the modified shear mold were fixed on a wooden bench with its optical axis normal to the studied surface, and the proper lighting to the tested specimen was applied. The setup of the experimental work is illustrated in Fig. 3. The modified shear mold was subjected to a vertical load in correspondence with the conventional direct shear test. The captured images of the examined sample are recorded before and during motion and/or deformation. In the center of the box, a transparent area was also kept in order to monitor the behavior of the soil sample when it is introduced into the box as shown in Fig. 3b. The area of interest was then selected on the investigated surface as highlighted in Fig. 4. During deformation, every portion of the pattern (i.e., subsets) is matched between the initial (reference) image and the deformed images, while the software is able to calculate the relative displacements. The studied image series were processed by using the open-source software DICe (SNL, Albuquerque, USA) that can be directly run on most operating systems.

Fig. 3 Experimental setup a test bench for direct shear, and b modified shear box

8

G. Alhakim et al.

Fig. 4 Area of interest monitored using digital image correlation and representation of the subset size

2.4 Validation of Results DIC results can vary widely among different setups, speckled patterns, matching parameters, etc. Therefore, it is important to use validation methods to check the obtained measurements. Two main methods were employed here in order to evaluate the DIC errors. On the one hand, the noise floor was calculated by taking and processing a series of 100 “stationary” (undeformed) images of the box to get an estimate of the threshold value of displacements under which measurements are indistinguishable from noise. On the other hand, the measured displacements between DIC and the transducers were compared. The induced controlled displacements to the shear boxes were achieved in this case by turning a screw that pushes the upper shear box, while the lower one is fixed.

3 Results and Discussion 3.1 Noise Floor Analysis The main source of random error in digital image correlation is camera noise. This is the temporal fluctuations of the gray levels that each pixel of the camera sensor observing a fixed object will perceive. The quantification of uncertainty considering the variance errors (or noise) is critical for rational assessment of DIC results. In this case, to study the noise floor, a series of 100 pre-test images for the static box were captured. Without any movement in the setup, the predictable displacement computed by DIC analysis is zero, and any measured displacement is noise. Eight

Analyzing the Potential Application of Low-Cost Digital Image …

9

Fig. 5 Selected check points on the upper and lower shear boxes

points were chosen on the monitored surface of the direct shear box within the desired field of view as displayed in Fig. 5. The results of noise in both X and Y directions are presented in Fig. 6. As it can be noticed, the average peak-to-peak ranges for all the selected points, or in other words the average difference between the lowest and highest detected displacement, are 0.0035 and 0.0052 mm for X and Y displacements, respectively. The standard deviation is generally computed to quantify the variance errors, and typically it should be similar for both X and Y directions. In this case, the standard deviation attained the values of 0.001 and 0.0012 in the horizontal (X) and vertical (Y) directions, respectively. It is worth mentioning that considering the maximum or average standard deviation between both directions, indicates that the values tend to be close to the mean of the set with a small amount of dispersion, thus, a low noise level. Moreover, the standard deviation was calculated for the check points selected on the plexiglass and the plastic mold distinctly. Particularly, points 0, 1, 2, and 3 correspond to the shear mold while points 4, 5, 6, and 7 refer to the plexiglass as shown in Fig. 5. In the first place, the standard deviation of the noise X-displacement reached 0.0011 and 0.0008 mm of the points selected on the mold and the glass, respectively. In addition, for the noise Y-displacement, the standard deviation of the chosen points attained 0.0013 for the mold, and 0.0011 for the glass. It can be noted that the standard deviation of the noise of the points selected on the mold surface is slightly greater than that of the glass in both directions. This could probably be due to the different planar positions of both materials, more precisely, the normal distance between the camera lens and the mold surface is relatively less than that of the glass plane.

10

G. Alhakim et al.

Fig. 6 Noise baseline in 8 checkpoints for a X-displacement, and b Y-displacement

3.2 Accuracy Assessment Figure 7 shows the displacements of the upper shear box measured by the extensometer from one side and processed by DIC method from the other side. To analyze the potential bias that could be introduced by the plexiglass, DIC results were also obtained for the glass and the plastic mold separately. To track these displacements, four points were selected on the speckled surface of each material, and it can be noticed that the results are in good agreement. Furthermore, Fig. 8 illustrates the relation between the displacements determined by the transducer and image processing technique. The correlation coefficients displayed on the graph depict a high fit between the results from DIC and LVDT for both the glass and the shear mold. The scatter plots visualizing bivariate data conform to a linear, strong relationship between both variables. The average estimated error between the results is about 0.1 mm. Thus, it can be concluded that the application of DIC in direct shear test is a reliable technique to study the mechanical behavior of soils particularly the shear strength and parameters. For a better illustration of the results, Table 1 presented the links for videos of the previously discussed results, namely, the noise floor analysis and the DIC measured displacements, in the horizontal and vertical directions, as well as the

Analyzing the Potential Application of Low-Cost Digital Image … Fig. 7 Displacements versus time steps

11

6 Shear box (DIC) Glass (DIC)

5 Displacement (mm)

Transducer 4 3 2 1 0 0

Fig. 8 Comparison of displacements measured by transducer and DIC

10

20 Time steps

30

40

6

Transducer displacement (mm)

Glass Shear box

5 4 3

y = 0.9708x - 0.0511 R² = 0.9991

2

y = 1.0219x - 0.1642 R² = 0.998

1 0 0

1

2 3 4 DIC displacement (mm)

5

6

component measurements. For instance, Fig. 9 showed a screenshot of the horizontal displacement captured at a certain time interval and processed by DIC technique.

12

G. Alhakim et al.

Table 1 Video links showing DIC results for noise floor and displacements Noise floor (stationary images)

Displacement measured by DIC (moving shear box)

Description

Video links

Absolute noise

https://doi.org/10.6084/m9.fig share.20154491.v1

Horizontal noise

https://doi.org/10.6084/m9.fig share.20154494.v1

Vertical noise

https://doi.org/10.6084/m9.fig share.20154488.v1

XY displacement component

https://doi.org/10.6084/m9.fig share.20154479.v1

Horizontal displacement

https://doi.org/10.6084/m9.fig share.20154482.v1

Vertical displacement

https://doi.org/10.6084/m9.fig share.20154485.v1

Fig. 9 Screen capture for X-displacement measured by DIC

4 Conclusions The results of the present research reveal the significance of applying the DIC technique on a properly prepared direct shear laboratory setup with the use of a modified shear box. The DIC application creates new possibilities for the effective and accurate analysis of deformations in soil materials when subjected to shear loading. And consequently, it compensates for the limitation of the standard direct shear test, particularly by measuring the non-uniform strain field and displacement within the soil sample rather than assuming a uniform distribution on the failure plane. The calculations of the noise floor of the setup and speckled pattern in terms of the mean and distribution of the displacements showed validated results with STD of 0.001 and 0.0012 mm in horizontal and vertical directions, respectively. Moreover,

Analyzing the Potential Application of Low-Cost Digital Image …

13

the displacements of the shear box, that was subjected to a lateral force introduced by turning a screw, measured by DIC are in conformity with the results of transducer deformations with an average error of 0.1 mm. The displacements were also measured separately for the plexiglass and the plastic box, and the results showed good agreement. Acknowledgements This work was supported by the EU Commission under the program Erasmus+ KA107 and Xunta de Galicia (Spain) under the grant “Financial aid for the consolidation and structure of competitive units of investigation in the universities of the University Galician System (2020-22)” Ref. ED341B 2020/25.

References 1. Dong YL, Pan B (2017) A review of speckle pattern fabrication and assessment for digital image correlation. Exp Mech 57:1161–1181. https://doi.org/10.1007/s11340-017-0283-1 2. Blaber J, Adair B, Antoniou A (2015) Ncorr: open-source 2D digital image correlation matlab software. Exp Mech 55:1105–1122. https://doi.org/10.1007/s11340-015-0009-1 3. Belloni V, Ravanelli R, Nascetti A, Di Rita M, Mattei D, Crespi M (2019) Py2dic: a new free and open source software for displacement and strain measurements in the field of experimental mechanics. Sensors (Switzerland) 19:1–19. https://doi.org/10.3390/s19183832 4. Arza-García M, Nuñez-Temes C, Lorenzana JA, Ortiz-Sanz JP, Castro A, Portela-Barral M, Gil-Docampo M, Bastos G (2022) Evaluation of a low—cost approach to 2—D digital image correlation vs . a commercial stereo—DIC system in Brazilian testing of soil specimens. Arch Civ Mech Eng 22:1–13. https://doi.org/10.1007/s43452-021-00325-0 5. Reu PL (2013) Uncertainty quantification for 3D digital image correlation. Conf Proc Soc Exp Mech Ser 3:311–317. https://doi.org/10.1007/978-1-4614-4235-6_43 6. Dematteis N, Giordan D (2021) Comparison of digital image correlation methods and the impact of noise in geoscience applications. Remote Sens 13:1–25. https://doi.org/10.3390/rs1 3020327 7. Huang Y, He X, Wang Q, Xiao J (2019) Deformation field and crack analyses of concrete using digital image correlation method. Front Struct. Civ Eng 13:1183–1199. https://doi.org/ 10.1007/s11709-019-0545-3 8. Fayyad TM, Lees JM (2014) Application of digital image correlation to reinforced concrete fracture, procedia. Mater Sci 3:1585–1590. https://doi.org/10.1016/j.mspro.2014.06.256 9. Mahal M, Blanksvärd T, Täljsten B, Sas G (2015) Using digital image correlation to evaluate fatigue behavior of strengthened reinforced concrete beams. Eng Struct 105:277–288. https:// doi.org/10.1016/j.engstruct.2015.10.017 10. Romeo E (2013) Two-dimensional digital image correlation for asphalt mixture characterisation: interest and limitations. Road Mater Pavement Des 14:747–763. https://doi.org/10.1080/ 14680629.2013.815128 11. Núñez-Temes C, Bastos G, Arza-García M, Castro A, Lorenzana Fernández JA, Ortiz-Sanz J, Portela M, Gil-Docampo M, Prego FJ (2022) Assessment of pavement deflection under vehicle loads using a 3D-DIC system in the field. Sci Rep 12:1–15. https://doi.org/10.1038/s41598022-13176-3 12. Górszczyk J, Malicki K, Zych T (2019) Application of digital image correlation (DIC) method for road material testing, Materials (Basel) 12. https://doi.org/10.3390/ma12152349 13. Eichhorn GN, Bowman A, Haigh SK, Stanier S (2020) Low-cost digital image correlation and strain measurement for geotechnical applications. Strain 56. https://doi.org/10.1111/str.12348

14

G. Alhakim et al.

14. Plé O, Tourabi A, Abuaisha MS (2013) 3-Dimensional digital image correlation for strains determination in clayey soil. Appl Mech Mater 353–354:463–466. https://doi.org/10.4028/ www.scientific.net/AMM.353-356.463 15. Liu W, Cao L, Li X, Li Y, Wang P (2011) Experimental study of silty clay’s deformation fields based on the principles of oedometer tests. Adv Mater Res 261–263:1539–1543. https://doi. org/10.4028/www.scientific.net/AMR.261-263.1539 16. Ko Y, Seo S, Jin T, Chung M (2021) Feasibility evaluation of the 3D-DIC non contact measurement system using small-scaled model test of earth retaining wall. Int J Geo-Eng 12. https:// doi.org/10.1186/s40703-021-00141-8 17. Tong BC (2022) Application of digital image correlations (DIC) technique on geotechnical reduced-scale model tests. J Korean Geosynth Soc 21:33–48. https://doi.org/10.12814/jkgss. 2022.21.1.033 18. Dirgelien˙e N, Skuodis Š, Griguseviˇcius A (2017) Experimental and numerical analysis of direct shear test. Procedia Eng 172:218–225. https://doi.org/10.1016/j.proeng.2017.02.052 19. Zhang L, Thornton C (2007) A numerical examination of the direct shear test. Geotechnique 57:343–354. https://doi.org/10.1680/geot.2007.57.4.343 20. Yan WM (2009) Fabric evolution in a numerical direct shear test. Comput Geotech 36:597–603. https://doi.org/10.1016/j.compgeo.2008.09.007 21. Shen J, Wang X, Liu W, Zhang P, Zhu C, Wang X (2020) Experimental study on mesoscopic shear behavior of calcareous sand material with digital imaging approach. Adv Civ Eng 2020. https://doi.org/10.1155/2020/8881264 22. Kong L, Chen F, Hua L (2014) Meso-direct-shear test of sands based on the digital image method. In: American society of civil engineers (ASCE), pp 265–274. https://doi.org/10.1061/ 9780784413388.027 23. IDICS, a good practices guide for digital image correlation. Int Digit Image Correl Soc 94. 10.32720/idics/gpg.ed1.%0Ahttp://idics.org/guide/

Applications of GIS in Estimating the Probable Maximum Earthquake Magnitude for Amaravati Region, Andhra Pradesh, India M. Madhusudhan Reddy , R. Siddhardha , G. Kalyan Kumar, and R. Suresh

Abstract An earthquake magnitude and their sizes have a significant impact on damage at a particular site. Hazard analysis provides the insights for vulnerability statistics of the area that is considered under study. Several important parameters including probable maximum earthquake magnitudes (Mmax ) are required to quantify the seismic risk. In this study to estimate the Mmax an area of 500 km radial distance has been considered as a seismic study area keeping a center latitude of 16° 52' N and longitude of 80° 51' E. An earthquake catalogue prepared for the period of 221 years, the earthquake data collected from 1801 to 2022, and the historical earthquake data has been collected from several data sources in order to investigate the seismicity of the area. The declustering techniques are used to remove the dependent events from earthquake data. All different earthquake magnitudes are converted into a moment magnitude (Mw ) using empirical relations. To estimate the seismic hazard parameters the catalogue completeness analysis was carried out; for this entire catalogue grouped into six different magnitude ranges with a constant bin width of 0.5 Mw. The whole extended study region has been separated into four source zones and designated with zone 1, 2, 3, and 4 based on estimated hazard parameters. The Mmax was estimated considering the data from each zone using Gupta and Kijko-Sellevoll-Bayes (KSB) methods. It has been observed that the obtained Mmax is varying from 6.1 to 6.9 Mw according to Gupta’s method whereas as per the KSB method the estimated Mmax varies from 6.0 to 7.7 Mw . The subzone-2 is contributing the highest value of Mmax compared to other zones. The estimated Mmax can be used further to quantify the seismic hazard risk of the Amaravati region. Keywords Earthquake catalogue · Maximum magnitude (Mmax ) · The incremental increase method · Kijko-Sellevoll-Bayes method M. M. Reddy (B) · R. Suresh Civil Engineering Department, Institute of Aeronautical Engineering, Dundigal, Hyderabad, India e-mail: [email protected] R. Siddhardha · G. K. Kumar Civil Engineering Department, National Institute of Technology Warangal, Hanamkonda, Telangana, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_2

15

16

M. M. Reddy et al.

1 Introduction The risks that come along with earthquakes almost always result in considerable damage to the buildings, and in some cases, it may even be the cause of fatalities that occur as a direct result of earthquakes in different parts of the globe. In recent decades, a number of significant earthquakes have struck India, causing enormous damage throughout the nation [1], even in the country’s stable continental region. According to the seismic microzonation map of India (2011), the whole of the nation has been separated into four distinct seismic zones, which have been designated as Zones II, III, IV, and V accordingly [2]. The seismic zonation map of India is given in Fig. 1, in which zones V and IV are classed as extremely high-risk and high-risk zones, whereas zones III and II are classified as moderate-risk to low-risk zones. On the other hand, the 2019 Earthquake Disaster Risk Index Report (EDRIR) states that a total of 59% of Indian lands are designated as being sensitive to earthquakes, and the intensity of the earthquakes would fall somewhere in the range of moderate to catastrophic. An experimental investigation was carried out by the National Disaster Management Authority (NDMA), India, and the International Institute of Information Technology, Hyderabad to study the risk index of 50 cities across the India and one complete district [3]. The cities that have been selected for evaluation risk are primarily based on the three major factors which include population density, housing threat factor followed by the smart cities as a factor. Initially, the risk associated with the selected cities has been categorised into low, medium, and high based on the seismic hazard map of India and other past research studies. The past few major earthquakes are actually considered as a strong evident to understand the influence of the local geological, geomorphological and other regional factors on damage during the earthquake. Some of major earthquakes that occurs across India are the Manipur (6.7 Mw ) in 2016, Nepal (7.8 Mw ) in 2015, Sikkim (6.9 Mw ) in 2011, Kashmir (7.6 Mw ) in 2005, Bhuj (7.7 Mw ) in 2001, Chamoli (6.8 Mw ) in 1999, and Jabalpur (5.8 Mw ) in 1997. After experiencing several major earthquakes the various governing bodies which are associated with the Indian government have made a decision to investigate subsoil stratification and identifying the earthquake prone areas across the country is very much needed and considered as most important and hence concluded that the outcomes of the planned investigations are effectively used as a guiding tool for land users for safe construction practises [4]. It has been observed by many researchers that the microzonation study will provide a best solution to estimate and control the seismic hazard during the earthquakes. Many researchers across the country are working towards making the microzonation procedure simple and realistic. The primary step in microzonation studies is the characterization of the local site conditions and estimation of seismic hazard analysis. In the current study, after making earthquake catalogue free from dependency events, the seismic hazard parameters such as a-value, b-value, and magnitude completeness (Mc ), β and Mmax have been estimated for the selected study region of area 217.3 km2 , where Mmax is nothing but it is the maximum magnitude of an earthquake that might be take place in the future. Thus, estimation of the expected Mmax is very important for parameters

Applications of GIS in Estimating the Probable Maximum Earthquake …

17

in seismic hazard analysis and hence these estimated Mmax values are significantly used to obtain the safe design criteria by the structural engineers. The estimated Mmax values are very commonly used by agencies like disaster management, builders, and investors to make sure the proposed structure to be safe throughout their life time [5, 6].

Fig. 1 The seismic zonation map of India (Source IS 1893: part (1)—2016)

18

M. M. Reddy et al.

2 Details of the Study Area In this current study, the expected Mmax for Amaravati region of Andhra Pradesh is evaluated by collecting earthquake data from the control region of 500 kms from Velagapudi as furnished in Fig. 2 and this extended boundary is considered as the seismic study area (SSA). The SSA is covers Telangana, Andhra Pradesh, Odisha, Chhattisgarh, Maharashtra, and Karnataka states of India. Among the six states four states covered partially and two states (Telangana, Andhra Pradesh) covered completely. All the necessary details of previous earthquakes have been collected from multiple data source centres which include the United States Geological Survey, the Indian Meteorological Department, the Amateur Seismic Centre, and the Geological Survey of India, since the earthquake data is the prime source to characterise the seismicity of the study region. The details of an historical earthquakes are compiled from the published articles of Chandra (1977), Rao and Rao (1984), and Guha and Basu (1993) [7–9]. The histogram of the earthquake data shows that the latest events including minor events are completed more accurately than the historical events. The observed maximum earthquake magnitude is considered as the reference magnitude to estimate the expected Mmax as per the Gupta’s method.

3 Characterization of Seismicity Seismicity of the study region characterised by compiling the earthquake data from multiple sources and for the mentioned period a total of 634 events was collected from the SSA of 500 km area. As a result of the instability of an earthquake measurement scale, particularly in the instrumental period, the data on earthquakes are published using a variety of scales which includes body wave magnitude, surface wave magnitude, local magnitude, and moment magnitude (Mw ). In order to keep uniform in the earthquake magnitudes, all of these earthquakes with different magnitudes are converted into moment magnitudes. The earthquake magnitudes converted using the widely popular relationships across the world were presented by Scordilis (2006) [10] and Heaton et al. [11]. The magnitudes of the earthquakes that are usually associated with the pre-instrumental era are often notified using an intensity scale and magnitude assigned purely based on the damage level of that particular earthquake; to convert such type of magnitudes into Mw the relationship proposed by GutenbergRichter (1956) has been used [12]. The declustering procedure was used in order to make the main event from its chain of dependent events. After the declustering process, the catalogue was prepared with 386 main events that were less than 3.0 Mw . Most of the situations measuring the smaller size earthquake magnitudes neglected particularly before 1976 AD, because there were no accurate instruments that are specially used to measure smaller size magnitude earthquakes, hence it is concluded that the completeness period of the earthquakes magnitudes related to the instrumental era is well established compared to pre-instrumental data. Figure 3 presents

Applications of GIS in Estimating the Probable Maximum Earthquake …

19

Fig. 2 Selected seismic study area

the histogram of the seismicity decay wise, which was developed to understand the pattern of seismicity over the selected study region. Further, the data accordingly decay wise has been used to identify the completeness period of the earthquake data for all different magnitude ranges; for this the methods of CUVI and Steeps (1972) have been used [13, 14].

20

M. M. Reddy et al.

Fig. 3 Histogram showing the seismicity pattern within the seismic study region decay wise

4 Estimation of Maximum Magnitude (Mmax ) Evaluation of Mmax is one of the essential criteria in seismic hazard analysis since there exist factors that are used to predict the degree of ground shaking that may occur during earthquakes. Also, it is utilised as an essential input parameter while computing the seismic hazard analysis. It has been observed from the seismicity pattern that the earthquake data is not evenly distributed over the region and in order to understand the pattern of seismicity intensity wise, four subzones were created using GIS software and by importing the earlier created shapefiles related to seismic sources a seismotectonic map of the Amaravati region has been generated. The Mmax is computed using two different scenarios. In the first scenario, all the events in the catalogue have been taken into account. Whereas, in the second scenario, the seismic study area of 500 km (Fig. 4) has been divided into four subzones (zone 1, zone II, zone III, and zone IV) based on the group of seismicity; catalogue completeness analysis, and seismic hazard parameters are estimated for the considered earthquake’s scenarios separately. Figure 4 presents the distribution of an earthquake’s magnitude

Applications of GIS in Estimating the Probable Maximum Earthquake …

21

over the region, divisions of subzones of the region under investigation and this map also illustrates the orientation of various seismic sources such as faults, lineaments, and shear zones. Next, for the purpose of conducting additional research, the catalogue has been roughly segmented into six magnitude group ranges, each bin with a constant width of magnitude 0.5 Mw beginning with 3.0 Mw . This is due to the fact that earthquakes with magnitudes that are lower than 3.0 Mw will, in most of the cases, not cause any damage to the structure. The histogram of all of the incidents broken down by zone can be seen in Fig. 5.

Fig. 4 Seismotectonic map of the Amaravati region including considered subzones

22

M. M. Reddy et al.

Fig. 5 Histogram of earthquake magnitudes related to four subzones

4.1 Gupta (2002) (Magnitude Incremental Method) According to the method suggested by Gupta’s (2002) an increment of 0.5 Mw is to be obs of that particular magnitude range, added to the observed maximum magnitude Mmax after adding the 0.5 Mw thar will be the expected maximum Mmax of the same magnitude range of the earthquake group[15]. Likewise, for all different magnitude ranges of the earthquakes groups an increment of 0.5 Mw is to be added for estimating the Mmax of each zone. This method is widely popular and preferred by many researchers across the world (Wheeler 2009) due to the wide range of users. The expected Mmax values for both scenarios according to Gupta’s method are summarized in Table 1. Table 1 Maximum magnitude by Gupta and KSB procedure Categories

Gupta method Observed Mmax

Mmax by KSB procedure Mmax

500 km area

6.4

6.9

6.52 ± 0.28

Zone 1

5.6

6.1

5.67 ± 0.26

Zone 2

6.4

6.9

7.70 ± 1.32

Zone 3

5.7

6.2

6.02 ± 0.40

Zone 4

5.6

6.1

0.90

Applications of GIS in Estimating the Probable Maximum Earthquake …

23

4.2 Kijko-Sellevoll-Bayes (KSB) Method Further, the Mmax is estimated as per the KSB method proposed by Kijko and Sellevoll [16]. According to the Kijko, the probable maximum earthquake of that region is called as Mmax of that respective zone. Due to the inconsistency in magnitude scales, uncertainties are involved in both the pre-instrumental and instrumental data. The Kijko procedure will consider the complete and incomplete earthquake catalogue during the estimation of Mmax using the following equation. Mmax = Mobs max +

) ] δ 1/q+2 exp[n · r q /1 − r q ] [ ( ⎡ −1/q, δ · r q − ⎡(−1/q, δ) (1) β

obs where, Mmax = observed maximum magnitude, β = 2.303b, δ = nCβ, n = number of earthquake events, r = p/(p + mmax- mmin), Cβ = normalizing coefficient of βvalue, q = ( β 2 ), σβ = standard deviation of β- value describes its uncertainty, β = (σβ ) mean value of β, p=( β 2 ) and ⎡(.,.) = Complementary incomplete gamma function. ( σβ ) The other seismic hazard parameters like b- value and λ—value was also estimated for both scenarios using MATLAB software HA3 code. The b-value varies from 0.35 to 0.67 and the range of λ—value 0.14 ± 0.03 to 1.66 ± 0.41. The Mmax values according to the KSB procedure as given in Table 1.

5 Conclusions In this study, an estimation of the possible maximum (Mmax ) earthquake magnitude for the Amaravati region was carried out using the incremental method that was proposed by Gupta (2002) and the Kijko-Sellevoll-Bayes (KSB) method for two different scenarios. These methods were used to calculate the magnitude of an earthquake using the Kijko-Sellevoll-Bayes (KSB) method. These two approaches were contrasted and compared in order to ascertain which one yielded the most precise findings. According to the findings, it has been determined that the incremental technique and the K-S-B method both state that the greatest earthquake took place in the subzone 2 (6.9 Mw according to the incremental method and 7.70 ± 1.32 Mw according to the K-S-B method). This conclusion was reached as a result of the findings. There have been a total of seventy earthquakes registered in Subzone 2, three of which have had magnitudes greater than 5.5. In addition to this, and particular information about a variety of seismic sources, including faults, lineaments, and shear zones, is shown in Fig. 4. It has been found that the Gundla–Kamma fault, the Addanki–Nujividu fault, and the Karempudi–Nakirekallu lineament are the most prominent active faults in this area. This conclusion was reached by determining which faults had the shortest distance between the event and the seismic source. In

24

M. M. Reddy et al.

addition, subzone 2 is home to both the Wajrakarur fault and the Kumadavati–Narihalla fault, and all of these faults can be found in subzone 2. In the future, both probabilistic and deterministic evaluations of the seismic hazard will make use of the value that was determined for Mmax. These analyses will be carried out in the future.

References 1. Reddy M, Konda RR, Kumar GK, Asadi SS (2020) Site characterization and evaluation of seismic sources for amaravati region. Int J Geotechn Earthq Eng (IJGEE) 11(1):71–86 2. Indian Standard, IS 1893 (Part I), 2016. Criteria for earthquake resistance design of structures, Part-I, Bureau of Indian Standard, New Delhi 3. Earthquake Disaster Risk Index Report 50 Towns and 1 Distrcit in Seismic Zone III, IV and V, 2019. Ministry of Home Affairs Government of India, NDMA Bhawan A-1, New Delhi 4. Anbazhagan P, Sitharam TG (2008) Seismic microzonation of Bangalore, India. J Earth. Syst Sci 117(S2):833–852 5. Anbazhagan P, Ketan B, Satyajit P (2015) Seismic hazard maps and spectrum for Patna considering region-specific seismotectonic parameters. Nat Hazards 78(2):1163–1195 6. Raghukanth STG (2011) Seismicity parameters for important urban agglomerations in India. Bull Earthq Eng 9(5):1361–1386 7. Chandra U (1977) Earthquakes of peninsular India-a seismotectonic study. Bull Seismol Soc Am 67(5):1387–1413 8. Ramalingeswara RB, Sitapathi RP (1984) Historical seismicity of Peninsular India. Bull Seismol Soc Am 74(6):2519–2533 9. Guha SK, Basu PC (1993) Catalogue of earthquakes (=> M 3.0) in peninsular India (No. AERB-TD-CSE--1). Atomic Energy Regulatory Board 10. Scordilis EM (2006) Empirical global relations converting Ms and mb to moment magnitude. J Seismol 10(2):225–236 11. Heaton TH, Tajima F, Mori AW (1986) Estimating ground motions using recorded accelerograms. Surv Geophys 8(1):25–83 12. Gutenberg B, Richter CF (1956) Earthquake magnitude, intensity, energy, and acceleration (second paper). Bull Seismol Soc Amer 46(2):105–145 13. Mulargia F, Tinti S (1985) Seismic sample areas defined from incomplete catalogues: an application to the Italian territory. Phys Earth Planet Inter 40(4):273–300 14. Stepp JC (1972) Analysis of completeness of the earthquake sample in the Puget Sound area and its effect on statistical estimates of earthquake hazard. In: Proceedings of the 1st international conference on microzonazion, Seattle, 2, 897–910 15. Gupta ID (2002) The state-of-the-art in seismic hazard analysis. ISET J Earthq Technol 39(4):311–346 16. Kijko A, Sellevoll MA (1992) Estimation of earthquake hazard parameters from incomplete data files, Part II, Incorporation of magnitude heterogeneity. Bull Seismol Soc Am 82(1):120– 134

Assessing the Effect of Land Use Land Cover Change on the Water Quality Index of a River Basin Using GIS and Remote Sensing Techniques W. S. Adhima, J. S. Gouri, Pooja N. Raj, P. S. Riya, and Lini R. Chandran

Abstract Water pollution is a major issue faced in both developed and developing countries. Land use land cover changes in urbanization, industrialization, and agricultural processes tend to have negative impacts on water quality at all scales. The water quality index serves as a commonly employed instrument for addressing data organization issues and assessing the effectiveness of management approaches aimed at enhancing water quality. The objectives of this study are estimation of water quality index (WQI) from water quality parameters for the study site using various methods and in determining the influence of land use land cover (LULC) change in water quality for the site. Using WQI, the classification of water quality can be done. GIS and remote sensing technology is used for estimating the LULC change for two time periods. The water quality parameters for calculating the WQI are collected from the Kerala State Pollution Control Board. Estimating WQI and identifying the relationship between LULC and water quality is important for effective and sustainable surface water quality management especially in reducing the pollutant concentration in water body. Keywords Water quality · Land use land cover · GIS · Remote sensing

1 Introduction Earth is covered with 75% of water, of which 96.5% is ocean, 2.5% is freshwater, and about 1% is salty water present in the ground. Freshwater sources are nearly 70% locked up in ice and the rest lies is in ground. Only 1.3% is surficial freshwater found mostly in lakes. Water Quality Index (WQI) comprises the data from different water quality parameters into a mathematical equation that evaluates and classifies the quality of water body. WQI was estimated using different methods with the aim to give a single value to the water quality of a source along with reducing higher number W. S. Adhima · J. S. Gouri · P. N. Raj · P. S. Riya · L. R. Chandran (B) Govt. Engineering College, Barton Hill, Trivandrum, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_3

25

26

W. S. Adhima et al.

of parameters [1]. The quality of water resources has been seriously deteriorating over years as a result of the Land Use Land Cover (LULC) change. Urban settlement areas and industries have a serious impact on water quality parameters. There was significant negative spatial relationship between water quality parameters and land use types [2]. This study estimates the WQI of a river basin in South Kerala using selected water quality parameters of the basin. This river basin serves as the major source of drinking water for Thiruvananthapuram, one of the metropolitanized cities located in the south western seacoast of the state of Kerala. The study enables us to: (1) assess the quality of surface water by establishing a WQI using different methods of its estimation (2) analyse the impacts of contaminants on water sources (3) establish the relationship between LULC change and water quality which leads to alteration of hydrological system. Finally, these results can provide a valid reference for the land use optimization and control of water pollution.

1.1 Objectives The objectives of the study are: (1) To evaluate the surface water quality of intake points of Karamana river using water quality parameters. (2) Developing WQI from the selected parameters using suitable methods. (3) Utilize GIS and Remote Sensing (RS) techniques to determine variations in LULC. (4) To assess the relationship between water quality and land use land cover changes in the Karamana River Basin (KRB).

1.2 Study Area The study area chosen was the Karamana river basin with the sample collecting stations located at Peppara, Aruvikkara, Mangattukadavu and Thiruvallam. It hosts the water supply to the Thiruvananthapuram city with its reservoirs at Peppara and Aruvikkara. Karamana River is a west flowing river through the city of Thiruvananthapuram in Kerala, India. The river originates near the southern end of the Western Ghats at Agastyar Koodam. The location map of the study area is given in Fig. 1 which was prepared using the watershed map of Karamana river prepared using GIS.

1.3 Water Quality Index Water Quality Index is a widely used tool to solve the problems of data management and to assess success and failures in management strategies for improving water quality. WQI was first developed by Horton in 1965 as an index rating system for

Assessing the Effect of Land Use Land Cover Change on the Water …

27

Fig. 1 Location Map of the study area

representing the water quality. WQI is a dimensionless number that incorporates various water quality parameters into a single number by normalizing the values and thus enabling easy interpretation of evaluating data. Estimating WQI includes three main steps: (1) selection of parameters (2) defining quality function for each parameter (3) aggregation through mathematical equation [3]. There are various methods to estimate WQIs according to the varying types and number of water quality parameters chosen. Some of the commonly used WQI methods are Weighted Arithmetic WQI (WAWQI) method, Nemerow’s Pollution Index (NPI) method for water quality assessment, National Sanitation Foundation WQI (NSFWQI) and Canadian Council of Ministers of the Environment WQI (CCMEWQI). The WQI of the present study area was estimated using WAWQI and NPI methods. Weighted Arithmetic Water Quality Index Method. In WAWQI method, the water quality status is ascertained using water quality rating and grading proposed by Brown et al. [4], and Chatterji and Raziuddin [12]. According to the Weighted Arithmetic Water Quality Index method, WQI is calculated as: W QI =

n Σ Q i Wi Wi i=1

(1)

where Qi = quality rating of ith water quality parameter, Wi = unit weight of ith water quality parameter. Qi describes value of the water quality parameter in polluted

28

W. S. Adhima et al.

water to standard permissible value, which is obtained as: (

vi − vio Q i = 100 si − vio

) (2)

where vi = estimated value of the ith parameter, vi0 = ideal value of the ith parameter, si = standard permissible value of the ith parameter. In majority of cases, vi /= 0 except for pH and Dissolved Oxygen (DO). For pH, vi0 = 7; for DO, vi0 = 14.6 mg/ l. The unit weight (Wi ) which is inversely proportional to values of the recommended standards is obtained as: k si

(3)

1 n Σ 1

(4)

Wi = where, k=

i=1

si

The rating of the water quality using WAWQI method is shown in Table 1 Nemerow’s Pollution Index (NPI) method. The NPI is one of the simplest WQI evaluation method compared to other methods. Dawood [5] successfully accessed the water quality for drinking and irrigation in five monitoring stations of Basrah marshes in the south of Iraq using the NPI. It is mathematically expressed as: N PI =

Ci Li

(5)

where Ci is the revealed concentration of ith parameter and Li is the allowable limit of allowable parameter. The unit of Ci and Li must be the same. The NPI value represents the general pollution provided by a single parameter. NPI has no units. When the NPI value exceeds 1, it indicates contamination of water. The parameters vary according to the study region. Table 1 Water Quality Rating and their Grading according to WQI values [4]

WQI

Water quality rating

Grading

0–25

Excellent

A

26–50

Good

B

51–75

Poor

C

76–100

Very poor

D

> 100

Non-potable

E

Assessing the Effect of Land Use Land Cover Change on the Water …

29

1.4 LULC Mapping Using GIS and Remote Sensing LULC refers to the categorization or classification of human activities and natural factors on the landscape within a specific time frame based on established statistical and scientific methods of analysis of source materials. It largely depends upon the ecological conditions, altitudes, geological structure and slope. In addition to the above factors, increase in population density, technological and other socio-economic factors also affect the LULC pattern. LULC studies tend to explain (1) what land cover types are changing, (2) where change is occurring, (3) the types of alteration occurring, (4) the rates or amounts of land change, and (5) the driving forces and proximate causes of change [6]. LULC mapping is enormously important for analysis of urban development and consequent land degradation. Satellite-based Remote Sensing, by virtue of its ability to provide comprehensive information regarding land use and land cover at a particular location and time, has significant impacts in the study of LULC. GIS help to integrate the acquired spatial and attribute data and to analyze the changes. There exist different image classification algorithms based on their rationality and complexity. Among those, Supervised classification is the most widely used method. In this method, the user can choose sample pixels in an image that are representative of specific classes and then direct the image processing software that is instructed to use these training sites as references for the classification of all other pixels in the image. Training data sets or input classes are chosen based on the user’s knowledge. The ERDAS Imagine software is used for image processing such as creating the False Colour Composite (FCC) and ArcGIS software is used for the supervised classification and to prepare land use land cover map and to determine the area under various features. The image classification can be done using multiple remote sensing features such as spatial, spectral, multi-sensor and multi-temporal. The numbers of LULC classes are preferred based on the requirement of a specific project for a particular application. The combination of water quality assessment method and spatial analysis tool in ArcGIS software greatly enhances the visualization of the research results. Remote Sensing and GIS provide an effective tool for LULC changes investigation. However, LULC changes based exclusively on Remote Sensing and GIS may not be reliable for particular environmental application at a local level.

30

W. S. Adhima et al.

2 Methodology 2.1 General The study area chosen is Karamana river basin with the sample collecting stations located at Peppara, Aruvikkara, Mangattukadavu and Thiruvallam. Karamana River is a west flowing river through the city of Thiruvananthapuram in Kerala, India. The river originates near the southern tip of the Western Ghats at Agastyar Koodam. It hosts the water supply to Thiruvananthapuram city with its reservoirs at Peppara and Aruvikkara.

2.2 Overview of Procedure The water quality analysis data for the stations were collected from the Kerala State Pollution Control Board (KSPCB) for the years 2012, 2018, 2019 and 2020. The water data regarding various parameters of surface water compiled by the River Rejuvenation Committee (RRC) of KSPCB was obtained to characterize the water quality. From the water quality data, the WQI was calculated using the WAWQI and NPI methods for the years 2012, 2018, 2019 and 2020. The parameters selected for the calculation of NPI are pH, EC, DO, TC and BOD. All parameters except TC is selected for WAWQI calculation. In the WAWQI method, after computing the WQI for the stations, the water quality status is ascertained using water quality rating and grading proposed by Brown et al. [4] and Chatterji and Raziuddin [12]. Meanwhile, in the NPI method the water quality parameters which have exceeded its standard limits are recognizable. The boundary of Karamana river basin was collected from Centre for Environment and Development Archive. The LULC changes in KRB was estimated from an input of Landsat 7 and Landsat 8 remote sensing images (30m resolution) for the years 2001 and 2020 respectively through the USGS Earth Explorer. False Color Composite (FCC) images are created by combining band 4 (NIR), band 3 (Red) and band 2 (Green) in Landsat 7, and band 5 (NIR), band 4 (Red) and band 3 (Green) in Landsat 8. Further, supervised image classification was done by ERDAS IMAGINE IV software and was reclassified in ArcGIS 10.2. The accuracy of the classified results was evaluated with the corresponding Google Earth images. The analysis of the results obtained using the above methodology helps in establishing a qualitative relationship between LULC and water quality.

Assessing the Effect of Land Use Land Cover Change on the Water …

31

3 Results and Discussion 3.1 Estimation of WQI The WQI was calculated using WAWQI and NPI method. The parameters selected for the calculation of WAWQI were: pH, EC, DO, and Biochemical Oxygen Demand (BOD). While for the calculation of NPI the selected WQ parameters were pH, EC, DO, BOD and TC. The selected years for WQI estimation are 2012, 2018, 2019 and 2020. The WAWQI calculations are given in the Tables 2, 3, 4 and 5. It was observed that the WQI of all the stations were consistently increasing from 2012 to 2019 with WQ rating of Peppara and Aruvikkara being ‘Good’ with a B grading throughout the years. This can be attributed to its location relatively at the upstream side of the basin. The water quality of Mangattukadavu, located just before entering the city, slightly deteriorated over the years. Thiruvallam, being the atmost downstream station and located within the corporation limits has a WQI value exceeding 100 for all the chosen years and thus falls under ‘non-potable’ category Table 2 WAWQI Calculations for 2012 Stations Peppara

Parameter

Vi

Si

Wi

Qi

WQI 35.64

pH

7.09

6.5–8.5

0.23

6

EC (µS/cm)

15

250

0.01

6

DO (mg/l)

7.37

5

0.38

75.31

BOD (mg/l)

0.7

5

0.38

14

1.00 Aruvikkara

pH

6.6

6.5–8.5

0.23

26.67

EC (µS/cm)

28

250

0.01

11.2

DO (mg/l)

6.8

5

0.38

81.25

BOD (mg/l)

0.8

5

0.38

16

pH

7.42

6.5–8.5

0.23

28

EC (µS/cm)

51.92

250

0.01

20.77

43.38

1.00 Mangattukadavu

DO (mg/l)

6.84

5

0.38

80.83

BOD (mg/l)

0.49

5

0.38

9.8

41.22

1.00 Thiruvallam

pH

7.38

6.5–8.5

0.23

25.33

EC (µ S/cm)

1008.92

250

0.01

403.57

DO (mg/l)

4.25

5

0.38

107.81

BOD (mg/l)

6.71

5

0.38

134.2

1.00

101.59

32

W. S. Adhima et al.

Table 3 WAWQI Calculations for 2018 Stations

Parameter

Vi

Si

Wi

Peppara

pH

6.9

6.5–8.5

0.23

Qi

WQI 6.67

EC (µS/cm)

28

250

0.01

11.20

DO (mg/l)

6.9

5

0.38

80.21

BOD (mg/l)

1.3

5

0.38

26.00

pH

6.7

6.5–8.5

0.23

20.00

EC (µS/cm)

102

250

0.01

40.80

42.31

1.00 Aruvikkara

DO (mg/l)

7

5

0.38

79.17

BOD (mg/l)

1.70

5

0.38

34.00

48.21

1.00 Mangattukadavu

pH

6.6

6.5–8.5

0.23

26.67

EC (µS/cm)

95.33

250

0.01

38.13

DO (mg/l)

6.7

5

0.38

82.29

BOD (mg/l)

1.60

5

0.38

32.00

50.13

1.00 Thiruvallam

pH

6.5

6.5–8.5

0.23

33.33

EC (µ S/cm)

225

250

0.01

90.00

DO (mg/l)

2.80

5

0.38

122.92

BOD (mg/l)

8.90

5

0.38

178.00

123.58

1.00

with an E grading. However, in 2020, there was a slight decline in the WQI values of all the stations indicating an improvement of water quality. The NPI calculations are shown in Tables 6 and 7. The results from NPI calculations specify water quality in terms of parameter concentrations. The NPI values of pH, EC and BOD were less than unity, indicating their concentrations confine within the standard limits. It was found that for all the chosen years, the NPI values of DO and TC exceeded unity inferring that they transcended their standard limits. At Thiruvallam, the NPI value of TC was severely high which implies further need for immediate treatment. In 2020, the NPI values decreased considerably for all the parameters at all the stations.

3.2 LULC Supervised Classification The LULC analysis for the years 2001 and 2020 was done with the help of ArcGIS and the classes for this analysis are forest, vegetation, built-up, water body and open area. The supervised classification provides the results of land use variations of these

Assessing the Effect of Land Use Land Cover Change on the Water …

33

Table 4 WAWQI Calculations for 2019 Stations

Parameter

Vi

Si

Wi

Peppara

pH

6.68

6.5–8.5

0.23

Qi 21.67

WQI

EC (µS/cm)

70

250

0.01

28.00

DO (mg/l)

6.35

5

0.38

85.94

BOD (mg/l)

0.6

5

0.38

12.00

pH

6.9

6.5–8.5

0.23

6.67

EC (µS/cm)

67

250

0.01

26.80

DO (mg/l)

6.43

5

0.38

85.16

BOD (mg/l)

0.93

5

0.38

18.50

58.51

1.00 Aruvikkara

82.32

1.00 Mangattukadavu

pH

6.9

6.5–8.5

0.23

6.67

EC (µS/cm)

94.75

250

0.01

37.90

DO (mg/l)

4.6

5

0.38

104.17

BOD (mg/l)

1.85

5

0.38

37.00

65.34

1.00 Thiruvallam

pH

7.03

6.5–8.5

0.23

1.67

EC (µ S/cm)

337.5

250

0.01

135.00

DO (mg/l)

2.35

5

0.38

127.60

BOD (mg/l)

7.5

5

0.38

150.00

124.86

1.00

years. The LULC classification of KRB for the year 2001 is conveyed in Fig. 2 and its corresponding pictorial representation is shown in Fig. 3. Figure 4 represents the LULC classification of KRB for the year 2020 and its corresponding pictorial representation is conveyed in Fig. 5 The percentage of change in area according to various LULC classes is shown in Table 8. It was found that the forest area depleted by one percent, i.e., from 17 percent during 2001 to 16 percent during 2020 leading to negative 8.7 percent change in area. The same trend was observed for vegetation whose percentage change in area was negative 29.33. Such unusual change was seen in the case of water body as its area increased by 8.27 percent. This is related to the COVID pandemic effect that resulted in lockdown. Researches conducted by Chakraborty et al. [7] and Aswathy et al. [8] proved the positive influence of lockdown in the case of water quality. Hence, the water demand drastically decreased due to shut down of industrial and commercial establishments as well as the disposal to river bodies came to rest for a couple of months leading to slight increase in quality and quantity. In case of open areas, the percentage change in area showed an increase of 22.65% with 5.66 km2 in 2001 to 6.94 km2 in 2020. It is obvious that the reduction in forests as well as vegetation will lead to increase in the built-up area. The LULC results backed up this

34

W. S. Adhima et al.

Table 5 WAWQI Calculations for 2020 Stations

Parameter

Vi

Si

Wi

Peppara

pH

6.68

6.5–8.5

0.23

Qi 21.67

WQI

EC (µS/cm)

70

250

0.01

28.00

DO (mg/l)

6.35

5

0.38

85.94

BOD (mg/l)

0.6

5

0.38

12.00

pH

6.9

6.5–8.5

0.23

6.67

EC (µS/cm)

67

250

0.01

26.80

DO (mg/l)

6.43

5

0.38

85.16

BOD (mg/l)

0.93

5

0.38

18.50

42.65

1.00 Aruvikkara

41.45

1.00 Mangattukadavu

pH

6.9

6.5–8.5

0.23

6.67

EC (µS/cm)

94.75

250

0.01

37.90

DO (mg/l)

4.6

5

0.38

104.17

BOD (mg/l)

1.85

5

0.38

37.00

55.92

1.00 Thiruvallam

pH

7.03

6.5–8.5

0.23

1.67

EC (µ S/cm)

337.5

250

0.01

135.00

DO (mg/l)

2.35

5

0.38

127.60

BOD (mg/l)

7.5

5

0.38

150.00

107.84

1.00

assumption because the percentage change in area for built-up was about 186.96. The built-up area was found to have upsurged from 81.14 km2 to 232.82 km2 .This results in Urban Heat Islands in which cities replace natural land cover with dense pavements, buildings and other heat absorbing surfaces. This makes the water in rivers, lakes, ponds and streams around the cities to be warmer than usual. Warm water holds less DO than cool water. This may be the cause of depletion of DO at Thiruvallam. Huang et al. [9] had analyzed the influence of LULC changes on the water quality parameters of the Chaohu lake basin and it was found that the built-up area and WQI were negatively related. Similarly, Permatasari et al. [10] analysed the LULC changes in Ciliwung watershed and found that the forest dominated and urban dominated areas had affected the water quality in a positive and negative manner respectively. The population growth of Thiruvananthapuram with respect to Kerala and India is indicated in Fig. 6. From this figure, it is inferred that the population of Thiruvananthapuram is increasing and according to an article published by The Hindu, certain studies showed that Kerala’s population will see an increase from 3.34 crores to 3.69 crores during 2011 to 2036.The population and the proportionate need for proper shelter leads to deforestation and clearing of vegetation for built up areas. These

Assessing the Effect of Land Use Land Cover Change on the Water …

35

Table 6 NPI Calculations for 2012 and 2020 2012 Stations Peppara

Aruvikkara

Mangattu-kadavu

Thiruvallam

Parameters

Ci

2018 Li

NPI

Ci

Li

NPI

pH

7.09

8.5

0.83

6.9

8.5

0.81

EC (µS/cm)

15

250

0.06

28

250

0.11

DO (mg/l)

7.37

5

1.47

6.9

5

1.38

BOD (mg/l)

0.7

5

0.14

1.3

5

0.26

TC (cfu/100ml)

293

50

5.86

100

50

2.00

pH

6.6

8.5

0.78

6.7

8.5

0.79

EC (µS/cm)

28

250

0.11

102

250

0.41

DO (mg/l)

6.8

5

1.36

7

5

1.40

BOD (mg/l)

0.8

5

0.16

1.7

5

0.34

TC (cfu/100ml)

500

50

10

110

50

2.20

pH

7.42

8.5

0.87

6.6

8.5

0.78

EC (µS/cm)

51.92

250

0.21

95.33

250

0.38

DO (mg/l)

6.84

5

1.37

6.7

5

1.34

BOD (mg/l)

0.49

5

0.1

1.6

5

0.32

TC (cfu/100ml)

77

50

1.54

2500

50

50.0

pH

7.38

8.5

0.87

6.5

8.5

0.76

EC (µS/cm)

1008.92

250

4.04

225

250

0.90

DO (mg/l)

4.25

5

0.85

2.8

5

0.56

BOD (mg/l)

6.71

5

1.34

8.9

5

1.78

TC (cfu/100ml)

11,075

50

221.50

52,000

50

1040

results go hand in hand with the population studies. In a recent study by Prasood et al. [11] it was found that the built-up area expanded progressively from 2001.

36

W. S. Adhima et al.

Table 7 NPI Calculations for 2019 and 2020 2019

2020

Stations

Parameters

Ci

Li

NPI

Ci

Li

Peppara

pH

6.66

8.5

0.78

6.675

8.5

0.79

EC (µS/cm)

64

250

5

70

250

0.28

DO (mg/l)

6.06

5

1.21

6.35

5

1.27

BOD (mg/l)

2.49

5

0.50

0.6

5

0.12

TC (cfu/100ml)

283.33

50

5.67

300

50

6.00

pH

6.57

8.5

0.77

6.9

8.5

0.81

EC (µS/cm)

31.92

250

0.13

67

250

0.27

DO (mg/l)

0.72

5

0.14

6.425

5

1.29

BOD (mg/l)

2.65

5

0.14

0.925

5

0.19

TC (cfu/100ml)

325.83

50

6.52

450

50

9.00

pH

6.48

8.5

0.76

6.9

8.5

0.81

EC (µS/cm)

71.35

250

0.29

94.75

250

0.38

DO (mg/l)

5.8

5

1.16

4.6

5

0.92

BOD (mg/l)

2.89

5

0.58

1.85

5

TC (cfu/100ml)

1233.33

50

24.67

1525

50

pH

6.45

8.5

0.76

7

8.5

0.82

EC (µS/cm)

321.67

250

1.29

337.5

250

1.35

DO (mg/l)

1.14

5

0.23

2.35

5

0.47

BOD (mg/l)

8.06

5

1.61

7.50

5

TC (cfu/100ml)

5608.33

50

112.2

4050

50

Aruvikkara

Mangattu-kadavu

Thiruvallam

NPI

0.37 30.5

1.50 81.0

Assessing the Effect of Land Use Land Cover Change on the Water …

Fig. 2 LULC Supervised classification of KRB (2001)

Fig. 3 Pictorial representation of LULC for the year 2001

37

38

Fig. 4 LULC Supervised classification of KRB (2020)

Fig. 5 Pictorial representation of LULC for the year 2020

W. S. Adhima et al.

Assessing the Effect of Land Use Land Cover Change on the Water …

39

Table 8 LULC Changes of KRB during 2001 and 2020 Year

2001

LULC classes

Area (km2 )

Vegetation

488.23

69.54

345.03

49.14

−29.33

81.14

11.56

232.84

33.16

186.96

7.46

1.06

1.15

8.27

Built-up area Water body Forest Open area Total

2020 Area (%)

Area (km2 )

8.07

% Change in area Area (%)

15.55

−8.70

5.66

0.81

6.94

0.99

22.65

702.10

100

702.10

100

119.62

17.04

109.21

Fig. 6 Population growth of Thiruvananthapuram with respect to Kerala and India

4 Conclusion The main objective of the present research was to establish a correlation between the land use land cover change and water quality. A qualitative negative relationship was inferred between them. Water quality Index clearly reveals the effect of urbanization in the river. Samples collected from Thiruvallam, which lies in the corporation limits shows maximum level of pollution in terms of BOD and a low DO content. Peppara reservoir which lies far off from the city and covered with its evergreen forests is not in danger of immediate pollution. Despite the various precautionary and action plans undertaken, the continuously declining water quality was clearly due to environmental stresses that resulted from rapid and dense urbanization and consequent increase in the built-up area over the observed span of years. There was a huge surge in the built-up area which expanded to nearly three times to that of 2001 within a span of two decades. According to various reports, the Thiruvananthapuram city, will be witnessing rigorous infrastructural development in the upcoming years which will result in further expansion of the

40

W. S. Adhima et al.

urban population. This will certainly catalyse the environmental deterioration. Also, localized pollution occurs from industrial units and slums located near the banks of the river. An observable enrichment of water quality was observed during 2020 owing to the reduction of industrial effluent discharge and shift in domestic and commercial water demand–supply patterns during the COVID-19 lockdown. However, restoration of all the activities post lockdown would result in the continuing trend of declining water quality. Thus, proper controlling measures are crucial to preserve the natural areas as well as to improve the quality of water. Policy makers and stakeholders must ensure checks for water quality at regular intervals of time. Action plans for reduction of net effluent discharge, regulation of channel precipitation and urban runoff must be enforced. The capacity of Sewage Treatment Plants must be increased proportionally with increase in population.

4.1 Scope of Future Work At present, there are diverse methods for the estimation of WQI. The absence of a universal WQI affects the uniformity of research works in the field of water quality. Thus, a well-established WQI which considers the vital WQ parameters and whose calculation is free from complex mathematical computations is to be formulated. It would be of great use if it can also be applied anywhere irrespective of the study area. Hence, more researches should be conducted in order to invent a suitable method for properly finding out the exact relationship. GIS proves to maximize the efficiency of decision making and planning in organizations and industries of all genres. This study has emphasized the application of Remote Sensing and GIS to develop LULC changes over the time for the river water quality based on the water quality parameters. LULC mapping has considerable importance in scientific, scholarly research, planning and management of water quality. GIS proves to be an effective means for data distribution and handling and helps to escalate the efficiency of planning and decision making. It has the ability to consolidate information from many sources and helps in the analysis of information involving geographical reference data at minimum cost.

References 1. Tyagi S, Sharma B, Singh P, Dobhal R (2013) Water quality assessment in terms of water quality index. Amer J Water Resour 1(03):34–38 2. Chen D, Elhadj A, Xu H, Xu X, Qiao Z (2020) A study on the relationship between land use change and water quality of the Mitidja watershed in Algeria based on GIS and RS. Sustainability 12(09):01–20 3. Uddin MG, Nash S, Olbert AI (2021) A review of water quality index models and their use for as-sessing surface water quality. Ecol Indicators 122(107218):1–21

Assessing the Effect of Land Use Land Cover Change on the Water …

41

4. Brown RM, McCleiland NI, Deininger RA, O’Connor MF (1972) A water quality index crossing the psychological barrier, indicators of environmental quality. In: 6th annual conference, advances in water pollution research, pp 787–794 5. Dawood AS (2017) Using of Nemerow’s Pollution Index (NPI) for water quality assessment of some Bas-rah Marshes, South of Iraq. J Babylon Univ/Eng Sci 25(05):1708–1720 6. Loveland TR, Acevedo W (2006) Land cover change in the Eastern United States, US Geological survey. https://landcovertrends.usgs.gov/east/regionalSummary.html. Accessed 28 June 2022 7. Chakraborty B, Bera B, Adhikary PP, Bhattacharjee S, Roy S, Saha S, Ghosh A, Sengupta D, Kumar P, Positive effects of COVID-19 lockdown on river water quality: evidence from River Damodar, India.https://doi.org/10.1038/s41598-021-99689-9 8. Aswathy TS, Francis S, Achu AL, Gopinath G (2021) Assessment of water quality in a tropical ram-sar wetland of southern India in the wake of COVID-19. Remote Sensing Appl Soc Environ 23(6) 9. Huang J, Zhan J, Yan H, Wu F, Deng X (2013) Evaluation of the impacts of land use on water quality. Scientific World J, 1–7 10. Permatasari PA, Setiawan Y, Khairiah RN, Effendi H (2017) The effect of land use change on water quality: a case study in Ciliwung watershed. In: IOP conference series: earth and environmental science, pp 1–7. IOP Publishing, Bogor, Indonesia 11. Prasood SP, Mukesh MV, Rani VR, Sajinkumar KS, Thrivikramji KP, Urbanization and its effects on water resources: scenario of a tropical river basin in South India. Remote Sen Appl Soc Environ.https://doi.org/10.1016/j.rsase.2021.100556 12. Chatterjee C, Raziuddin M (2002) Determination of water quality index of a degraded river in Asanol In-dustrial area, Raniganj, Burdwan, West Bengal. Nat Environ Pollut Technol 01(02):181–189

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground Water Levels in Kurukshetra, Haryana Vikas Singh and A. K. Prabhakar

Abstract Analyzing the trajectory of hydrological indicators is crucial due to significant reduction in water supplies, as well as the rise in demand. The aim of the current study is to evaluate the status of ground water levels in Kurukshetra district at all the seven Blocks viz. Thanesar, Shahbad, Pipli, Babain, Ladwa, Pehowa, Ismailabad for past one decade i.e., from 2010 to 2020. The current study examines ground water variation and trend analysis during the Pre-Monsoon and Post-Monsoon seasons. Also, Inversed Distance Weight (IDW) interpolation approach was used for creating ground water contour maps by using the Arc-GIS 10.8. Results depict that in the Pre-Monsoon study Ismailabad block reflected maximum fluctuation and Pehowa Block reflected minimum fluctuation. In the Post-monsoon study Pehowa Block reflected maximum fluctuation and Ladwa block reflected minimum fluctuation. For Pre-Monsoon, the rate of Groundwater depletion for one decade at seven blocks of Kurukshetra viz. Thanesar 38.88%, Shahbad 27.34%, Pipli 31.86%, Pehowa 6.44%, Ladwa 25.67%, Ismailabad 29.55%, Babain 29.52%. For the Post-Monsoon the rate of Groundwater depletion for one decade for seven blocks of Kurukshetra viz. Thanesar 46.97%, Shahbad 30.32%, Pipli 28.47%, Pehowa 51.58%, Ladwa 25.57%, Ismailabad 36.14%, Babain 27.97%. Average rate of Groundwater Depletion during Pre-Monsoon and during Post-Monsoon seasons for one decade in Kurukshetra district is 27.03%.and 35.28% respectively. As a result, regional entities must design a framework for sensible use of ground resources, with effort made to increase groundwater recharge in order to ensure enough groundwater in the future. Keywords Groundwater · Pre-Monsoon · Post-Monsoon · Fluctuations · GIS

V. Singh (B) · A. K. Prabhakar Civil Engineering Department, NIT Kurukshetra, Kurukshetra, Haryana, India e-mail: [email protected] A. K. Prabhakar e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_4

43

44

V. Singh and A. K. Prabhakar

1 Introduction 1.1 General Water is one of the most basic requirements of all living beings. It is essential for life, whether it is in the form of groundwater, surface water, or precipitation. It is impossible to imagine life on Earth without water [1]. Groundwater among the available sources plays a critical role due to demands for drinking water, irrigation, livestock water, and industrial activities. Currently, 90% of rural water supply, 50% urban water supply, and 70% agricultural water supply in India depend on groundwater trapped in porous and permeable layers of the soil, referred to as aquifers [2]. Aside from human demands, groundwater is required to maintain the biological water cycle since it serves as a source of recharging for lakes, river, and wetlands. Secular variation refers to changes in ground water level over a long period of time [6]. Long-term fluctuations in level are caused by alternating seasons of wet and dry years, in which rainfall is above and below the mean, respectively. The regulating factor is recharge, which is determined by rainfall intensity and dispersion, as well as the amount of surface runoff. A downward trend in ground water level may persist for many years in overdeveloped basins where draught exceeds recharge [3]. Seasonal variations caused by factors such as rainfall and irrigation, as well as pumping discharge, all of which follow well-defined seasonal cycles [4]. The beginning and end of the irrigation seasons are marked by the highest and lowest levels, which occur around April and September, respectively. Groundwater is chosen over surface water because it is more cost-effective, more convenient, and less polluted than surface water [5]. India is approaching a point when meeting the rising water demands of a geometrically growing and unevenly distributed population would be difficult. Rapid urbanization, caused by economic and social developments, has altered the consuming network and whole water requirement, affecting particularly places with hard and impervious terrain, where groundwater access is problematic [7]. The main problem is unrestricted access to groundwater, as well as waste during agricultural operations. Because of the higher consumption rate than the usual sluggish rate of groundwater recharge, the groundwater level is rapidly diminishing. Given the approaching shortage of available groundwater sources in the near future, it has become critical for water scientists and planners to quantify available water in order to make wise use of them [9].

1.2 Significance of Present Study • This study will be helpful in predicting trends in the groundwater depth with respect to the development of the area. • To suggest some preventive measures for the future decades to use the groundwater effectively will be the major scope of the research.

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground …

45

1.3 Objective of the Study • To analytically study the recent trends of the Groundwater Fluctuation levels in Kurukshetra district of Haryana that is majorly affected due to development. • To assess long-term impacts of the groundwater table depth and the fluctuations in Pre-Monsoon and Post-Monsoon conditions.

2 Study Area The Kurukshetra district is situated in the north section of Haryana and is surrounded by the Haryana districts of Ambala, Karnal, Yamunanagar, and Kaithal, as well as the Punjab districts of Patiala, with area geographically covering around 1530 km2 . Between 29°52' to 30°12' N latitude and 76°26' to 77°04' E longitude is the location of the District. The district’s major rivers are the Markanda and Saraswati. The Saraswati River was once considered to flow across Haryana, but it has since vanished. Between the Yamuna and the Sutlej, the Ghaggar rises in the lower Shivalik highlands of the outer Himalayas and reaches Haryana near Pinjore district Panchkula. It passes via Ambala and Hisar on its way to Bikaner in Rajasthan, where it runs for 290 kms before disappearing into the Rajasthan deserts. During the rainy season, it also affects a large portion of the Kurukshetra District. Kurukshetra district lacks a perennial river. After emerging from the Nahan hills, the river Markhanda flows through the district’s northwestern portion. Yamuna and Saraswati, two holy rivers, flow along the eastern and northern borders, respectively (Fig. 1).

3 Methodology 3.1 General The research work was carried out in Arc GIS 10.8 version at Siemens Centre of Excellence (SCOE) in Water Resources Engineering Lab of the Civil EngineeringDepartment, National Institute of Technology, Kurukshetra, Haryana, India. GIS software illustrates geographical information through layer building maps, such as temperature data or trade flows, and enables the handling and analysis of geographic information. Like the majority of GIS programs, ArcGIS produces maps with categories arranged as layers. Each layer is geographically recorded, allowing the computer to precisely align the layers when they are layered on top of one another to create a complex data map [8]. Almost always, the base layer is a geographic map that has been assembled from many sources depending on the type of presentation needed (satellite, road map, etc.). The first three layers are known as feature or vector

46

V. Singh and A. K. Prabhakar

Fig. 1 Location of the study area

layers, and each one contains distinct functions that are distinguished by the platform. These are points, lines, and polygon and raster images. Data can be mapped and analyses done using at least one of these spatial layers, and can be connected with factors like demographic changes or data tables (Fig. 2).

3.2 IDW Contours in ArcGIS IDW i.e., Inverse Distance Weighted Interpolation implicitly assumes things that are close to one another that have more in common than things that are far apart. IDW predicts a value for any unknown point using the measured data in the vicinity of the prediction location. The expected value is more affected by the measured values

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground …

47

Fig. 2 Steps involved in research work

that are in close proximity to the prediction location than by those that are farther away. According to IDW, each measured point has a local impact that gets smaller as you move away. It offers higher weights to points that are closer to the prediction location, and the weights decrease as the distance increases, therefore known as Inverse Distance Weighted [10]. The weights are determined by taking the inverse of the distances (between the data point and the projected point) raised to the power p. As a result, as the distance increases, the weights fall quickly. Depending on the value of p, the weights decrease at various rates. If p = 0, the forecast will be the mean of all the data values in the search region because each weight I is equal. As p increases, the weights for distant points decrease quickly. If the p value is really high, only the nearby surrounding points will have an impact on the projection. Geostatistical analysts employ power levels greater than or equal to one. When p = 2, the inverse distance squared weighted interpolation technique is applied. Although there is no theoretical rationale for selecting p = 2 above other values, the impact of changing p should be investigated by looking at the cross-validation statistics and the output preview. Because objects that are close to one another are more similar than those that are farther apart, the measured values will have little relevance to the value of the forecast location as the locations go farther apart. You can eliminate the more distant sites that will have little impact on the prediction to speed up calculations.

3.3 Box Plot In simple terms, the box plot can be defined in terms of descriptive statistics ideas. That is, a box or whiskers plot is a graphical representation of numerical data

48

V. Singh and A. K. Prabhakar

Fig. 3 Details of box plot

grouped through their quartiles. The words box-and-whisker plot and box-andwhisker diagram refer to the variability outside the lower and higher quartiles, which is indicated by lines extending from the boxes or whiskers. Individual points can be used to indicate outliers. Graphs can be used to determine how much the data values change or spread out. The box plot comes in handy when we need more information than only the measures of central tendency. It also takes up less space. It’s a form of pictorial depiction as well. Because the Centre, spread, and overall range are instantly visible, the distributions can be easily compared using these box plots (Fig. 3).

4 Result and Discussion 4.1 Groundwater Table Level Average in the Pre- Monsoon Season in 2010 Groundwater levels during the Pre -Monsoon period in the year 2010 varied from 26.08 m to 37.11 m below the ground level. Water Table was highest in theThanesar Block and lowest in the Pehowa Block of the district (Fig. 4).

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground …

49

Fig. 4 Pre-Monsoon 2010 average ground water level

4.2 Groundwater Table Level Average in the Pre- Monsoon Season in 2020 Groundwater levels during the Pre - Monsoon period in the year 2020 varied from 33.82 m to 45.22 m below the ground level. Water Table was lowest in Shahbad Block and highest in the Ladwa Block of the district (Fig. 5).

50

V. Singh and A. K. Prabhakar

Fig. 5 Pre-Monsoon 2020 average ground water level

4.3 Groundwater Table Level Average During Post Monsoon Period in 2010 Ground water levels during the Post-Monsoon period in the year 2010 varied from 25.65 m to 34.92 m below the ground level. Water Table was highest in the Thanesar and lowest in the Shahbad block of the district (Fig. 6).

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground …

51

Fig. 6 Post-Monsoon 2010 average ground water level

4.4 Groundwater Table Level Average During Post Monsoon Period in 2020 Ground water levels during the Post-Monsoon period in the year 2020 varied from 33.68 m to 45.51 m below the ground level. Water Table was highest in the Ladwa Block and lowest in the Shahbad block of the district (Fig. 7).

52

V. Singh and A. K. Prabhakar

Fig. 7 Post-Monsoon 2020 average ground water level

5 Conclusion For a better understanding, long-term Seasonal average groundwater levels were interpreted (Pre-Monsoon and Post-Monsoon), from the above done works, the following key conclusions were drawn: • In the study it was found that during the Pre-Monsoon Season over the decade 2020 the fluctuations were maximum at the Thol station in Ismailabad Block i.e., 10.25 m and the fluctuations were minimum at the Galedwa station in Pehowa

Assessment of Fluctuations in Pre-monsoon and Post-monsoon Ground …

53

Block i.e., 2.39 m. During the Post-Monsoon Period over the decade from 2010 to 2020 the fluctuations were maximum at the Galedwa station in Pehowa Block i.e., 13.87 m and the fluctuations were minimum at the Lohara station in Ladwa Block i.e., 6.86 m. • For the Pre-Monsoon period, the rate of Groundwater depletion for one decade at Thanesar block 38.88%, Shahbad block 27.34%, Pipli block 31.86%, Pehowa block 6.44%, Ladwa block 25.67%, Ismailabad block 29.55%, Babain block 29.52%. For the Post-Monsoon period, the rate of Groundwater depletion for one decade at Thanesar block is 46.97%, Shahbad block 30.32%, Pipli Block 28.47%, Pehowa block 51.58%, Ladwa block 25.57%, Ismailabad block 36.14%, and Babain block 27.97%. • Average rate of Groundwater Depletion during the Pre-Monsoon Period and during the Post-Monsoon Period for one decade in Kurukshetra district is 27.03%.and 35.28% respectively. • During Pre-Monsoon ground the southern and western areas of the district had higher water levels than the remaining of the district. In the Post-Monsoon Period groundwater table level was high in Thanesar and Ladwa blocks i.e., the southern and eastern parts of Kurukshetra. The analysis of water level patterns is crucial for determining the future scenario of accessible Groundwater resources of the state. The success of various groundwater recharge strategies could also be determined via trend analysis. As a result, regional entities must design a framework for sensible use of present ground resources, with every effort made to increase Groundwater recharge in order to ensure enough supplies in the future.

References 1. Aneja R (2017) Ground water level in Haryana: a challenge for sustainability. Int J Res Analyt Rev 4(3):43–48 2. Devi K, Singh M (2021) Evaluating ground water fluctuations in district Hisar, Haryana: a Temporal Analysis (1974–2018). PalArch’s J Archaeol Egypt/Egyptol 18(4):6985–6994 3. Duhan AK (2017) Groundwater pumping irrigation in Haryana: issues and challenges. Int J Res Geogr 3(2):18–21 4. Goyal SK, Chaudhary BS (2010) GIS based study of Spatial-Temporal changes in groundwater depth and quality in Kaithal district of Haryana, India. J Indian Geophys Union 14:75–87 5. Goyal SK, Chaudhary BS, Singh O, Sethi GK, Thakur PK (2010) Variability analysis of groundwater levels—AGIS-based case study. J Indian Soc Remote Sens 38(2):355–364 6. Indhulekha K, Chandra Mondal K, Jhariya DC (2019) Groundwater prospect mapping using remote sensing, GIS and resistivity survey techniques in ChhokraNala Raipur district, Chhattisgarh, India. J Water Supply Res Technol AQUA 68(7):595–606 7. Joshi SK, Gupta S, Sinha R, Densmore AL, Rai SP, Shekhar S, van Dijk WM (2021) Strongly heterogeneous patterns of groundwater depletion in Northwestern India. J Hydrol 598:126492 8. Krishan G, Sharma LM, Yadav BK, Ghosh NC (2016) Analysis of water level fluctuations and TDS variations in the groundwater at Mewat (Nuh) district, Haryana (India). Current World Environ 11(2):388

54

V. Singh and A. K. Prabhakar

9. Kumar PJ (2022) GIS-based mapping of water-level fluctuations (WLF) and its impact on groundwater in an Agrarian District in Tamil Nadu, India. Environ Dev Sustain 24(1):994–1009 10. Kumar S, Rani M, Verma S (2020) Assessment of pre-monsoon and monsoon groundwater level and fluctuation at regional level in south of Haryana; Reference to Rewari District. Int J Appl Sci Eng 8(2):123–129

Assessment of Land Use—Land Cover Changes in District Dehradun (1991–2021) Madhusudan Thapliyal and A. K. Prabhakar

Abstract The region of district Dehradun in Uttarakhand, India has shown a rapid growth in population and urbanization in the last few decades. An attempt was made to assess the land use-land cover (LULC) changes in the entire district in the last 30 years from 1991 to 2021. Satellite imagery from the Landsat—5, 8 with a spatial resolution of 30 m was used as the remotely sensed data and the open-source software QGIS was used to process the data and produce the LULC maps and other important statistics using the maximum likelihood algorithm. The study has shown that during 1991–2021, dense forest, active cropland and sparse vegetation areas have decreased by 17.52%, 62.11% and 7.50% respectively whereas built-up and bare soil areas have increased by 112.23% and 44.32% respectively. No definite conclusions regarding change in areas could be reached for water and dry river bed but in general, the total stream flow area was observed to be decreasing. The average overall accuracy of the classifications was found to be 85.01% with a kappa coefficient of 0.794. Keywords GIS · Landuse · Temporal and spatial LULC change analysis · Classification & accuracy assessment

1 Introduction Remote sensing is a very useful technique for detecting land cover changes in rapidly developing regions so as to provide useful data for landuse planning and policy making. Landuse changes are a serious concern for developing cities due to rapid increase in population which increases the requirement of land viz. housing, road, agricultural and water etc. and causes issues such as land degradation and problems in the water cycle [4, 7, 14]. Several others researchers [5, 6, 10, 15] found that changes M. Thapliyal (B) Climate Hydrology Division, National Institute of Hydrology, Roorkee, India e-mail: [email protected] A. K. Prabhakar Department of Civil Engineering, National Institute of Technology, Kurukshetra, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_5

55

56

M. Thapliyal and A. K. Prabhakar

Fig. 1 The study area (District—Dehradun, State—Uttarakhand, Country—India) (Right to Left)

in landuse such as decrease in forest area, increase in urbanization and agricultural area are adversely impacting the water balances. In the current study, proper methods have been adopted and land use-land cover maps have been generated for the region of district Dehradun in Uttarakhand for different time frames with satisfactory to good levels of accuracy and hence, spatio-temporal LULC change analysis has been done to analyze the changes in the landuse pattern.

1.1 The Study Area The entire district of Dehradun in Uttarakhand, India was chosen as the study area of the research. Dehradun lies in the extreme western part of its state Uttarakhand. Its geographical extent lies between 29° 58' N to 31° 02' N and 77° 34' E to 78° 19' E. The district has river Yamuna flowing at its west and river Ganga at its east. A large part of the district is under forest cover and its topography is a mixture of hills and valleys. The region is also known for its good climatic conditions. The study region can be seen in the following figures (Figs. 1 and 2).

1.2 Need for the Study The region of the Dehradun district has gone through a rapid urbanization in the last few decades due to the gradual migration of a large number of people from the hilly districts of the state (for various reasons like education, healthcare facilities and ease of living) as well as from other states. Over the years, there has been a lot of infrastructure development in terms of highway expansions, construction of buildings, residential colonies, commercial centres etc. which has led to the crucial loss of some agricultural lands and forest cover of the region. As per the literature

Assessment of Land Use—Land Cover Changes in District Dehradun …

57

Fig. 2 Elevation Map—District Dehradun

review done, the land use-land cover study for the entire district had not been done considering a long period of time. So, the study was found necessary and was carried out for the entire Dehradun district for a time period of 30 years from 1991 to 2021 in order to analyze the long-term changes in areas of various land cover features. The study might be useful for any further research work or atleast for the availability of generated LULC maps and statistics of the adopted study area.

58

M. Thapliyal and A. K. Prabhakar

2 Data and Methodology 2.1 Data Used The satellite imagery data was obtained from United States Geological Survey’s website earthexplorer.usgs.gov. The considered study area was a part of the region falling in the Worldwide Reference System (WRS-2)’s Path-146 and Row-39. All the images were dated between 15th Dec (of previous year) and 20th Jan (of the study year) and had cloud covers between 2 and 9%. The images were in the CRS—EPSG: 32644 (WGS-84 / UTM Zone-44N). For years 1991, 2001 and 2011, the image data was from the satellite ‘Landsat-5’ with the sensor ‘Thematic Mapper (TM)’. For year 2021, the data was from the satellite ‘Landsat-8’ with sensors ‘Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS)’. The satellite images used, had a spatial resolution of 30 m.

2.2 Methodology LULC maps have been generated for the region of district Dehradun for different time frames viz. 1991, 2001, 2011 and 2021. Atmospheric correction was not performed on the satellite images as the cloud cover for all images was less than 10% and was negligible over the study area. Although, the values of data sets were converted to spectral reflectance (in %) in order to obtain the spectral signature plots of various LC classes. The entire process of mapping the land use-land cover of the study area was performed in the open source software ‘QGIS Desktop-3.20.1’. The study region was extracted from the satellite imagery using ‘Clip Raster by Mask Layer’ function of QGIS. Then by using the ‘Semi-Automatic Classification Plugin’, FCCs were generated using different band combinations to identify various LC classes. For classifying, training samples (initially 15 to 20 in numbers) were provided for all LC classes. The classification was performed using ‘Land Cover Signature’ classification and by applying the ‘Maximum Likelihood’ classification algorithm for the unclassified pixels. More number of training samples were subsequently added to increase the mapping accuracy upto a good extent in terms of observation. In order to remove the pixels appearing in the form of noise due to misclassification, classification sieve provided in QGIS was used. The classified area and accuracy statistics were generated from QGIS itself and some raster-based area analysis was also done using QGIS’s ‘Raster Calculator’ and later the required statistical analysis and other interpretations were carried out using Microsoft Excel. The complete process has been shown in Fig. 3.

Assessment of Land Use—Land Cover Changes in District Dehradun …

59

Fig. 3 Process flowchart of the study

2.3 Accuracy Assessment For accuracy assessment, samples in the form of individual pixels were created (total 120–150 in numbers) from the classified map and then the accuracy was checked by comparing the sample pixel areas with google satellite images and the FCCs of the respective years. Stratified random sampling approach was used to take the samples from the LULC maps for accuracy assessment and for deciding the total number of samples, the following equation was used from Cochran [3]. ⎛ ⎞2 )2 Σ Wi S i W S i i ≈⎝ ( ) ⎠ n=[ ]2 ( / ) Σ ˆ 2 ˆ S O S(O) + 1 N Wi S i (Σ

where, ‘n’ is the total number of sample units, ‘N’ is the number of units in the ROI (total number of pixels), ‘Wi ’ is the mapped area √ proportion of class ‘i’, ‘Si ’ is the standard deviation in mapping class ‘i’, Si = Ui(1 − Ui), ‘Ui ’ is the conjectured value of user’s accuracy for class ‘i’ and ‘S(Ô)’ is the standard error in the estimated overall accuracy. Considering the standard error of overall accuracy as 0.03 (or 3%), the total number of samples was calculated. For distributing it among the various LC classes, three different sample size allocations were made [8] for all the land cover classes based on equal sample size for all the classes, mapped area proportion of the classes and standard deviation in mapping of classes. Then, the final number of samples for each class was calculated by simply taking average of the three allocations.

60

M. Thapliyal and A. K. Prabhakar

3 Results and Discussions The following results (shown in Table 1) have been obtained after the classification of satellite imageries of the considered study years viz. 1991, 2001, 2011 and 2021 for the district Dehradun: As the mapping of ‘Dry River Bed’ and ‘Water’ was not very accurate and also as they shared a low proportion of area in the district, leaving them, all other classes (forest, cropland, built-up, bare soil and sparse vegetation) were considered and referred to as ‘major’ land cover classes. In order to remove the snow cover from the maps, the 1991 map was compared with the 2001 map using the raster calculator and it was concluded that as per the increase/decrease of LC class proportions, 25%, 43% and 32% of the snow cover area should be replaced with forest, bare soil and sparse vegetation respectively. After doing the same, the finally estimated percentage areas for the major LC classes along with their rate of change have been given and shown below in Table 2 and Fig. 4. The results (as per Table 2) have shown that during 1991–2021, dense forest, active cropland and sparse vegetation areas have decreased by 17.52%, 62.11% and 7.50% respectively whereas built-up and bare soil areas have increased by 112.23% and 44.32% respectively. Area wise (as per Table 3 ahead), there has been a loss of Table 1 Year wise percentage areas of different LC classes S. No

LULC class type

1991

2001

2011

2021

1

Forest

40.036

39.401

37.012

33.937

2

Cropland

2.185

1.682

1.148

0.828

3

Built-up

2.421

3.109

4.367

5.138

4

Water

1.144

1.207

0.687

1.937

5

Dry river bed

3.324

2.322

1.744

2.092

6

Bare soil

17.410

22.095

25.693

27.891

7

Sparse/dry vegetation

28.997

30.147

29.349

28.137

8

Snow cover

4.482

0.038

0.000

0.040

Table 2 Finally estimated percentage areas of major LC classes 1991 LULC class type

2001

2011

2021

Rate of change (%)

Overall 1991–2001 2001–2011 2011–2021 change (%) (1991–2021) −6.09

−8.28

−17.52

Cropland

2.185

1.682

1.148

0.828 −23.04

−31.70

−27.90

−62.11

Built-up

2.421

3.109

4.367

5.138 +28.38

+40.50

+17.65

+112.23

19.337 22.111 25.693 27.908 +14.35

+16.20

+8.62

+44.32

Sparse 30.431 30.159 29.349 28.150 −0.89 vegetation

−2.69

−4.09

−7.50

Forest

Bare soil

41.157 39.411 37.012 33.947 −4.24

Assessment of Land Use—Land Cover Changes in District Dehradun …

61

Fig. 4 Change in percentage areas of major LC classes with the years

222.63 km2 , 41.9 km2 , and 70.45 km2 of forest, cropland and sparse vegetation areas respectively and the gain of 83.9 km2 and 264.68 km2 of built-up and bare soil areas respectively, considering the whole district of Dehradun. The following LULC maps have been generated for the considered time frames viz. 1991, 2001, 2011 and 2021 for the Dehradun district (Fig. 5) and the decadal changes in LULC features are given in Table 3. It can be inferred from the land cover classification results that the agricultural lands are decreasing at a rapid rate whereas the natural vegetation areas are also decreasing, but gradually. The built-up regions are increasing very fast due to the increase in migration and settlement of people into the district which is leading to the requirement of more construction. Bare soil regions also, have been found to increase. This is due to the fact that agricultural lands are getting converted into residential plots which remain vacant for a long time and also due to clearance of forests for road construction and other development works. From the results, the river area also seems to be decreasing probably due to the encroachment around the stream courses. Similar results have been reported in other studies [2, 9, 12, 13]. Using the raster calculator provided in QGIS, some areas were identified where urban/built-up areas were not present in 1991, but appeared in 2021. They can be seen in the map (Fig. 6).

939.72

Sparse vegetation

931.31

682.80

95.99

74.77

597.13

Built-up

51.93

1217.00

1270.91

67.47

2001

1991

Bare soil

Cropland

Forest

LULC class type

906.29

793.41

134.87

35.46

1142.92

2011

Table 3 Year wise (decadal) absolutes areas of LC classes

869.27

861.81

158.67

25.57

1048.28

2021

+110.62 −25.03

+85.66 −8.40

+38.87

−74.08 −16.46

−53.92 −15.54 +21.22

2001–2011

1991–2001

Change in area (km2 )

−37.02

+68.39

+23.80

−9.89

−94.64

2011–2021

−70.45

+264.68

+83.9

−41.9

−222.63

Overall change (km2 ) 1991–2021

62 M. Thapliyal and A. K. Prabhakar

Assessment of Land Use—Land Cover Changes in District Dehradun …

Fig. 5 Generated LULC Maps of District Dehradun

63

64

M. Thapliyal and A. K. Prabhakar

Fig. 6 Hotspots of urbanization

3.1 Problems Faced During Classification Following issues were faced during the classification of land use-land cover for the region of district Dehradun: • A large amount of area of the district in the northern part was mountainous having high elevation and slopes and therefore the satellite imageries of these areas contained large shadow regions which were difficult to map as they distorted (darkened) the spectral signature of the land cover features. Similarly, high regions facing the sensor brightened up and also contributed to the difficulty in classification. Also, topographic corrections through algorithms were unsuccessful in solving the problem due to high slopes. • Mapping of croplands and built-up areas in mountainous regions was very difficult due to topographic variations, shadows etc. as stated above. • In 1991, a significant amount of area was under snow cover which is temporary, so some analysis was required to know what was the actual land cover beneath it. • The river sand or ‘Dry River Bed’ had a spectral signature very similar to ‘BuiltUp’ areas and hence they both often got cross-classified. Large buildings with

Assessment of Land Use—Land Cover Changes in District Dehradun …

65

Table 4 Classification accuracies LULC class no

LULC class type

Average accuracy PA

Average Kappa Hat

UA

1

Forest

99.26

90.88

0.862

2

Cropland

53.71

82.20

0.818

3

Built-up

71.83

65.84

0.647

4

Water

64.84

37.52

0.368

5

Dry river bed

70.70

71.63

0.711

6

Bare soil

82.48

92.09

0.896

7

Sparse vegetation

8

Snow Overall

79.91

78.41

0.699

100.00

94.87

0.947

85.01

0.794

smooth finish or shiny material showed higher reflectance and were classified as dry river bed and some of the river sand was classified as built-up. • Also, the dark shadow areas often got misclassified as water. Therefore, the mapping of water (bodies) was not much accurate. Mostly, it was over-estimated. • As the spatial resolution of the satellite sensors was 30 m, so small features such as minor roads, small unclustered buildings, small fields etc. couldn’t be mapped accurately.

3.2 Classification Accuracy Inspite of the problems listed above, good levels of overall accuracy were reached as the problem was mostly with the classes having lesser percentage of area, which is reflected in the accuracy assessment results (Table 4). The accuracy of classification was overall 85.01% with a kappa coefficient of 0.794 which can be considered good taking into account the spatial resolution of the satellite data. It can be seen in Table 4, that for forest, and bare soil, the accuracy was good; for cropland, built-up areas, dry river bed and sparse vegetation, the accuracy was fair; for water, the accuracy was satisfactory to poor whereas, for snow cover, the accuracy was excellent.

3.3 Relation of LULC with Population As Dehradun district has experienced a population explosion for past decades, it was also tried to find the relation of land-use with the growing population. Table 5 shows the population stats from official census data.

66

M. Thapliyal and A. K. Prabhakar

Table 5 Population data—District Dehradun (as per official Census data)

Year

Population

1961

4,29,014

Growth Rate

1971

5,77,306

34.57

1981

7,61,668

31.93

1991

10,25,679

34.66

2001

12,82,143

25.00

2011

16,96,694

32.33

2021 (expected)

22,34,512

31.70

–

Table 6 Relation of LULC changes with population Parameter/Year

1991

2001

2011

2021

Correlation with population

Population

10,25,679

12,82,143

16,96,694

22,34,512

NA

Forest (km2 )

1270.91

1217.00

1142.92

1048.28

−0.9994

Sparse vegetation (km2 )

939.72

931.31

906.29

869.27

−0.9939

Cropland (km2 )

67.47

51.93

35.46

25.57

−0.9697

Built-up (km2 )

74.77

95.99

134.87

158.67

+0.9856

Bare soil (km2 )

597.13

682.80

793.41

861.81

+0.9785

It can be inferred from Table 6 that forest, sparse vegetation and croplands have a very high negative correlation with population whereas built-up area and bare soil have a very high positive correlation with the population of the district. It means that, with the growth of population, the natural vegetation cover and the agricultural lands are decreasing rapidly whereas the bare soil regions and built-up areas are increasing so as to fulfil the needs of space and infrastructure for the large number of people in the district.

3.4 Impact of LULC Changes There have been various studies over the impact of landuse changes. Considering the effect of urbanization, i.e. the increase in built-up, open areas and the decrease in agricultural, vegetation areas (as can be seen in this study), there can be two, very prominent impacts related to hydrology. One can be the increase in runoff due to increase in built-up/impermeable areas, and the other can be the decrease in groundwater recharge as more water will flow away as runoff. Similar changes have been observed in Dehradun. The study by Anh et al. [1] shows that due to the increase in built-up area by 88.65% in the Asan watershed (major changes in subbasins 20 and 16), and the decrease in cropland and forest areas by 6.61% and 0.25%,

Assessment of Land Use—Land Cover Changes in District Dehradun …

67

respectively, there was an average increase of 70% and 65% in the surface runoff from the above sub-basins during the 13 years’ time period due to the rapid growth of urban industries in the area. Also, the study by Shukla et al. [11] shows that the historic well hydrograph data for Dehradun city for the period 1997–2008, indicates that groundwater levels in the most of the wells have shown a declining trend. Such findings are indicative of the impact of land cover changes on the hydrology of a region.

4 Conclusion Remote Sensing and GIS are definitely very helpful ways of tracking the changes in land use-land cover for developing regions that need study over a large time frame. Areas with plain topography can be mapped with a good accuracy even with open source softwares and freely available satellite data. The study was quite successful in assessing the LULC changes in the study region of district Dehradun. Statistically, the study has shown that during 1991–2021, dense forest, active cropland and sparse vegetation areas have decreased by 17.52%, 62.11% and 7.50% respectively whereas built-up and bare soil areas have increased by 112.23% and 44.32% respectively. The results also show that a lot of natural vegetation cover of the district has been lost and the rate of decline of it is growing every decade. Another crucial observation was regarding the rapid loss of agricultural land due to building of residential colonies. The district is very well known for the production of various crops including the famous basmati rice, but it might soon lose that fame. Proper landuse planning is required to limit the urbanization so as to preserve the natural heritage of the district. Also, the people with farming lands should be encouraged and supported so that they can help to conserve the agricultural areas of the district. Landuse planning is also important to keep a check on the water resource availability as the increase of built-up/impermeable areas also seems to have an impact on the hydrological characteristics such as runoff and groundwater storage of the district.

References 1. Anh NN, Chouksey A, Aggarwal SP (2016) Assessment of land use/land cover change impact on the hydrology of Asan River Watershed of Dehradun District, Uttarakhand. Int J Current Eng Technol 6(4):1125–1131 2. Bhat PA, ul Shafiq M, Mir AA, Ahmed P (2017) Urban sprawl and its impact on landuse/land cover dynamics of Dehradun City, India. Int J Sustain Built Environ 6(2):513–521 3. Cochran WG (1977) Sampling techniques, 3rd edn. Wiley, New York 4. Farida & Noordwijk MV (2004) Analysis of river discharge due to land-use change and application of GenRiver model in Way Besai watershed, Sumberjaya. World Agroforestry CentreICRAF

68

M. Thapliyal and A. K. Prabhakar

5. Li C, Sun G, Caldwell PV, Cohen E, Fang Y, ... Meentemeyer RK (2020) Impacts of urbanization on watershed water balances across the conterminous United States. Water Resour Res 56(7):e2019WR026574 6. Mawardi I (2010) River Basin Watershed damage and decrease the carrying capacity of water resources in Java as well as efforts to handle. Hidrosfir Indonesia J 7. Nugroho P, Marsono D, Sudira P, Suryatmojo H (2013) Impact of land-use changes on water balance. Procedia Environ Sci 17:256–262 8. Olofsson P, Foody GM, Herold M, Stehman SV, Woodcock CE, Wulder MA (2014) Good practices for estimating area and assessing accuracy of land change. Remote Sens Environ 148:42–57 9. Patidar S, Sankhla V (2015) Change detection of Land–use and Land-cover of Dehradun City: a spatio-temporal analysis. Int J Adv Remote Sens GIS 4(1):1170–1180 10. Prabhakar A, Tiwari H (2015) Land use and land cover effect on groundwater storage. Model Earth Syst Environ 1:1–10 11. Shukla AK, Ojha CSP, Garg RD, Pal L, Satyavati SP (2018) Groundwater appraisal of Dehradun City, Uttarakhand, India using remote sensing and geographical information system technology. Conference: STIWM at IIT Roorkee, India 12. Thapa R, Bahuguna V (2021) Monitoring land encroachment and land use & land cover (LULC) change in the Pachhua Dun, Dehradun District using landsat images 1989 and 2020. JGISE: J Geospatial Inf Sci Eng 4(1):71–80 13. Tiwari K, Khanduri K (2011) Land use/land cover change detection in Doon valley (Dehradun tehsil), Uttarakhand: using GIS& Remote sensing technique. Int J Geomat Geosci 2(1):34–41 14. Turkelboom F, Poesen J, Trébuil G (2008) The multiple land degradation effects caused by landuse intensification in tropical steeplands: a catchment study from northern Thailand. CATENA 75(1):102–116 15. Wang Y (2020) Urban land and sustainable resource use: unpacking the countervailing effects of urbanization on water use in China, 1990–2014. Land Use Policy 90:104307

Comparison of Streamflow Simulations for Different DEMs Nagireddy Venkata Jayasimha Reddy and R. Arunkumar

Abstract Hydrological modeling of the rainfall-runoff process is crucial not only for understanding the characteristics of a basin but also for sustainable planning and management of water resources. Most of the hydrologic models require digital elevation model (DEM) as the prime input, from which the basin and its stream network are delineated by using GIS tools. DEMs from various satellite imageries like Cartosat, SRTM, ASTER, IKONOS, and others have been widely used for such purposes. However, the primary challenge is the selection of an appropriate source of DEM, since the accuracy produced from satellite imageries varies. This study aims to compare the streamflow simulations in the Chaliyar basin using a HECHMS model developed from two different DEMs. The Cartosat-DEM and ASTER-DEM are considered for comparison. Results of the study show that Cartosat-DEM gives a clear watershed boundary and area of 2912 km2 which is very close to the IndiaWRIS data, whereas the ASTER-DEM gives slightly better results in simulating the streamflow based on performance measures of the model. Keywords HEC-HMS · ASTER-DEM · Cartosat-DEM · Chaliyar basin

1 Introduction Hydrological models are simplified replication of the real world system [8]. They not only simulate various hydrological processes but also a better choice in predicting catastrophic events like floods and droughts [10]. In the previous two decades, hydrological models have evolved in a number of ways, from conceptual to physically based distributed models. Within a hydrological model, complexity varies from simple empirical models to complicated process-based numerical models [9]. Physically based hydrological models are developed based on complex hydrological processes, as well as regionally varying information regarding watershed topography N. V. J. Reddy (B) · R. Arunkumar Department of Civil Engineering, National Institute of Technology Calicut, Kozhikode 673601, Kerala, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_6

69

70

N. V. J. Reddy and R. Arunkumar

such as soil properties and land use patterns. Several tools like SWAT, HEC-HMS, VIC, SWMM, MIKE-SHE, etc. are available for developing hydrological models. However, the selection of the appropriate tool is governed by the basin characterstics and the objectives of the study [4]. Accurate simulation of streamflow in a basin is always challenging because of the complexity associated with the hydrological processes [6]. Prediction of streamflow is one of the most important tasks in hydrological modelling [5]. When it comes to hydrological assessments and streamflow forecasting, it’s crucial to use the proper model. However, the complexity lies in selecting an appropriate model and this aspect has received little attention. The hydrological model selection mainly depends on the data availability, spatial and temporal scales, model complexity, cost, convenience of use, and level of knowledge. Verma et al. [10] used the HEC-HMS model for simulating the streamflow in the upper Sabarmati basin, Gujarat. The model performance was very good with coefficient of determination and Nash–Sutcliffe Efficiency as 0.88 and 0.84, respectively. [1] estimated the peak flows and abstraction losses by integrating GIS with HECHMS and calibrated with the actual flow. It was reported the simulated flows were close to actual flows. From this, HEC-HMS is one of the most suitable rainfall run-off models to simulate streamflow for Indian conditions. DEM is the main input for the rainfall runoff model which can be taken from many sources. The primary objective of the study is to compare the simulation of streamflow using the HEC-HMS model developed from different DEMs. Thus, this study concentrates on comparison of Cartosat-DEM and ASTER-DEM that can be used as an input to the HMS model to generate streamflow.

1.1 Hydrologic Engineering Center-Hydrologic Modeling System (HEC-HMS) Model The HEC-HMS is a physically-based semi-distributed model developed by the US Army Corps of Engineers and is widely used to simulate a variety of hydrological processes [2]. Mostly the rainfall-runoff processes are modeled using the HMS. It is helpful for addressing a wide variety of challenges in a broad range of geographic situations. The hydrology of floods, small urban or natural watershed drainage, and the water supply from huge river basins are all included in this. Studies on water availability, urban drainage, flow forecasting, impacts of urbanisation, design of reservoir spillways, mitigation of flood damages, floodplain control, and system operation can all be studied using the HEC-HMS model, either on their own or in conjunction with other tools. HEC-HMS model has mainly three components, a basin model, meteorological model and control specifications. Basin model transforms the atmospheric conditions into streamflow in the watershed. To achieve this the watershed is separated into smaller unit is using hydrologic principles. These smaller units are linked together to

Comparison of Streamflow Simulations for Different DEMs

71

form a dendritic network in order to model the stream network [2]. HMS has various elements like junctions, sub-basins, sources, sinks, reaches, etc. through which the basin is represented in a spatial context. The meteorological model specifies the boundary conditions that have an impact on the watershed during a simulation. If the basin model includes sub-basins, then meteorological model must be specified for each sub-basin. Control specifications are one of the most important aspects of an HMS model as they control the duration of simulations and their beginning and termination. Runs of simulation are the primary method used to calculate results. One set of meteorological conditions, and one set of control requirements constitute each run. Using the results in HEC-HMS, graphs, summary tables, and time-series tables can be made using the basin map and the watershed explorer.

1.2 Digital Elevation Model (DEM) The three-dimensional space is represented by digital elevation models (DEMs) in a discrete form. DEM is useful for characterizing topography and determining basin boundaries as well as studying the terrain inside the watershed. DEM from satellite imageries such as ASTER, SRTM, Cartosat etc. have evolved over the years with various spatial and temporal resolutions and used for a wide range of applications. The DEM derived from Cartosat is widely used in India for many applications. ISRO launched Cartosat-1 in 2005 which renders stereo data for entire India. Using digital photogrammetric techniques, DEM (1 arc Sec) are produced for the entire country. Photogrammetric blocks were created to produce seamless DEMs, and DEMs were edited throughout the scenes with few break-lines. Utilizing ground control points, through dense feature matching, triangulated irregular network modelling, and automated strip mosaicking, seamless DEM data creation is carried out automatically. To identify distortions, the quality verification technique uses panning and draped visualization. Height accuracy for stereo pairs with overlapping sections is further validated. The other DEM that is widely used globally is the ASTER DEM, a 3A1 product that is obtained from the ERSDAC, Japan-based ASTER science team. This georeferenced satellite imagery includes visible, infrared, thermal, bands and a DEM. The nadir and backward-looking infrared bands are used to create the DEM. The spatial resolution of ASTER DEM is 30 m. DEMs serve a variety of functions and are necessary prerequisite for numerous applications. They are especially helpful in areas without comprehensive topographic maps. Geomorphometry, which focuses on surface processes like landslides and may be directly represented from a DEM, is one area of study where DEMs have been proven to be helpful [3].

72

N. V. J. Reddy and R. Arunkumar

2 Materials and Methods 2.1 Data The data required for the study includes observed daily precipitation which is collected from the IMD for 20 years from 1995 to 2014. DEM of 30 m resolution is obtained from NASA Earth data (ASTER) and Bhuvan (Cartosat). The observed daily streamflow for a period 1995 to 2014 is obtained from India-WRIS.

2.2 Methodology HEC-HMS is a physically-based semi-distributed model which is widely used to simulate the streamflow. The essential components are basin model, meteorological model and control specifications which can simulate the streamflow. The basin is delineated with the help of ArcGIS using both ASTER-DEM and Cartosat-DEM. The inputs for the HEC-HMS model are delineated watershed, daily observed precipitation, sub basin parameters and observed streamflow. In HMS many methods are available to calculate canopy loss, surface loss, infiltration loss, baseflow, transformation and routing. One component that may be applied to the sub-basin element and used to reflect the presence of vegetation in the landscape is the canopy. The amount of precipitation that reaches the ground surface is decreased by vegetation as it intercepts precipitation. Between storm events, intercepted water evaporates. Transpiration is the term for the process by which vegetation draws water from the soil. Evaporation and transpiration are frequently combined which is known as evapotranspiration. Although choosing a canopy approach is optional, it should be used for applications involving continuous simulation. HMS offers three different methods: simple canopy, gridded simple canopy, and dynamic canopy. Canopy loss calculations in this study are done using a simple canopy. One of the elements that can be a part of the sub-basin element is the surface. It represents the ground surface where water may accumulate in surface depression storage. Gridded simple surface and simple surface are the methods that are available in HMS. In this model, the simple surface method is employed. Although a sub-basin element conceptually represents infiltration, surface runoff, and subsurface processes, the actual infiltration calculations are performed by a loss method that is integrated into the sub-basin. There are total of twelve distinct loss methods offered, including deficit and constant, layered Green and Amptdeficit, exponential Green and Ampt, gridded Green and Ampt, gridded SCS curve number, initial and constant, etc. In this work, the infiltration loss is estimated using the deficit and constant approach. A baseflow algorithm integrated within the subbasin executes the real subsurface calculations. There are six different baseflow techniques available, including recession baseflow, linear reservoir baseflow, and

Comparison of Streamflow Simulations for Different DEMs

73

constant monthly baseflow. The recession method is used to determine baseflow in this study. Within the sub-basin, a transform approach is used to calculate the real surface runoff. Several unit hydrograph approaches, a kinematic wave implementation, a linear quasi-distributed method, and a two-dimensional (2D) diffusion wave method are some of the popular techniques. In this model, the excess precipitation is converted to runoff using a Clark unit hydrograph. Although a reach element conceptually represents a segment of stream or river, a routing method is placed within the reach actually performs the calculations. There are nine different routing options available. To route the reaches to sink, the Muskingum method is applied. The calculated streamflow is calibrated with the observed streamflow once the model parameters and control specification are provided.

2.3 Study Area The study area, as shown in Fig. 1, is the Chaliyar River Basin in Kerala located in India. With a length of 169 km, the Chaliyar is the fourth-longest river in Kerala and flows through the districts of Malappuram and Kozhikode. It starts 2066 m above sea level in the Elambalari Hills in Gudalur taluk of Nilgiris district in Tamil Nadu and drains into the Lakshadweep Sea at Beypore. Its coordinates are 75° 35' and 76° 45' East longitude and 11° 05' to 11° 40' North latitude. The area of the basin as a whole is 2933 km2 , of which 2545 km2 are in Kerala and the rest 388 km2 are in Tamil Nadu.

Chaliyar basin

Kerala

Legend Kilometers

Meteorological stations Kuniyil guage and discharge station Streams

Fig. 1 Location of Chaliyar basin

74

N. V. J. Reddy and R. Arunkumar

Fig. 2 Digital Elevation Model of Chaliyar basin

In the basin, the average annual rainfall is about 3012 mm. The southwest (June– September) and northeast (October–November) monsoons are the two main rainy seasons. Major thunderstorm activity is prevalent during the pre-monsoon months (March to May), whereas throughout the winter months (December-February). The four main soil types in the river basin, which account for 60.73%, 24.56%, 9.85%, and 4.86% of the total area, are clay, gravelly loam and loam, gravelly clay. The overall area is made up of urban areas, water bodies and rocky areas with agricultural land making up approximately 74.26 percent of the total area and forests making up 14.21 percent. There is one CWC hydrological observation station at Kuniyil and two IMD meteorological stations in the watershed located at Nilambur and Manjeri as shown in Fig. 1. Figure 2 shows the ASTER DEM with the elevation ranges from 0 to 2604 m.

3 Results and Discussions 3.1 Results HEC-HMS model calibration for streamflow is performed using observed daily streamflow data for eight years (1995–2002) and validated for 3 years (2011–2013). The RSR, NSE, R2 , PBIAS values of model simulations run during calibration are

Comparison of Streamflow Simulations for Different DEMs

75

evaluated; the estimated and observed streamflows are graphically assessed for agreement. The parameter values are modified until the observed streamflow and computed streamflow are relatively close to each other. Both manual and automatic calibration are performed. For automatic calibration, simplex and univariate methods are used. Both the Cartosat-DEM and ASTER-DEM final calibrated values for the model parameters are included in Table 1. The performance parameters adopted in the study are RMSE-observations standard deviation ratio (RSR), Nash–Sutcliffe efficiency (NSE), coefficient of determination (R2 ) and percentage bias (PBIAS). Moriasi et al. [7] criteria are used for model rating. The model performance during the calibration and validation are listed in Table 2. Cartosat-DEM and ASTER-DEM are compared with the observed streamflow at the Kuniyil for 1995–2002 to check how simulated and observed streamflow are closely related as shown in Fig. 3 and Fig. 4 respectively. From Fig. 3, it is clear that the day i.e., 12-07-1997 of simulated peak flow is matched with the observed peak flow and the peak flows are 1859.2 m3 /s, 2193 m3 / Table 1 Calibrated model parameters for both Cartosat-DEM and ASTER-DEM Model

Parameter

Canopy

Initial storage 14% Maximum storage (mm) Crop 10 coefficient 0.90

Loss

Initial deficit (mm) Maximum deficit (mm) Constant rate (mm/hr)

6.77 21.55 0.48

4.73 14 0.46

Transform

Time of concentration (hr) Storage coefficient (hr)

16.33 56

16.33 56

Baseflow

Initial discharge (m3/s) Recession constant Ratio to peak

19 0.90 0.42

17 0.90 0.42

Routing

K (hr) X

0.32 0.20

0.32 0.10

Calibrated value (Cartosat-DEM)

Calibrated value (ASTER-DEM) 16% 7 0.80

Table 2 Comparison of performance indices of simulated and observed streamflow during calibration and validation for Cartosat-DEM and ASTER-DEM Performance measures

Cartosat-DEM Calibration period

ASTER-DEM Validation period

Calibration period

Validation period

RSR

0.55

0.68

0.54

0.66

NSE

0.70

0.53

0.71

0.56

R2

0.58

0.59

0.58

0.59

−3.81

2.50

−5.92

2.76

PBIAS

76

N. V. J. Reddy and R. Arunkumar

Fig. 3 Comparison of simulated and observed daily streamflow for Cartosat-DEM

Fig. 4 Comparison of simulated and observed daily streamflow for ASTER-DEM

Comparison of Streamflow Simulations for Different DEMs

77

s respectively. The model predicts 15.22% lesser than the observed value when the peak flow is compared for the Cartosat-DEM. From Fig. 4, it is clear that the day i.e., 12-07-1997 of simulated peak flow is matched with the observed peak flow and the peak flows are 1739.9 m3 /s, 2193 m3 / s respectively. The model predicts 20.66% lesser than the observed value when the peak flow is compared for the ASTER-DEM.

3.2 Conclusion In this study, the streamflow simulations are compared for two DEMs. The streamflow is simulated through HEC-HMS model. During the calibration of HEC-HMS model, it is observed that time of concentration, initial deficit, maximum storage and constant rate are highly sensitive parameters as small change in the parameters affects the streamflow. From the results, it is noticed that the HEC-HMS model has simulated the streamflow close the observed values. The performance of the model is very good during calibration and good during validation for both DEMs. It is observed that the Cartosat-DEM and ASTER-DEM give the all most similar results. On comparing the performance measures of both DEMs R2 , NSE and RSR do not show significant change whereas the PBIAS has slight change. Cartosat-DEM gives the clear watershed boundary of 2912 km2 which is very close the India-WRIS, Chaliyar river basin boundary. Over all, comparing the performance measures of model, ASTER-DEM gives slightly better results over Cartosat-DEM.

References 1. Alfy M (2016) Assessing the impact of arid area urbanization on flash floods using GIS, remote sensing, and HEC-HMS rainfall–runoff modeling. Hydrol Res Nordic Assoc Hydrol 47(6):1142–1160 2. Feldman AD (2022) Hydrologic modeling system HEC-HMS user’s manual 3. Gajalakshmi K, Anantharama V (2015) Comparative study of cartosat-DEM and SRTM-DEM on elevation data and terrain elements. Cloud Publ Int J Adv Remote Sens GIS, 1361–1366 4. Halwatura D, Najim MM (2013) M, Application of the HEC-HMS model for runoff simulation in a tropical catchment. Environ Model Softw 46:155–162 5. Hamdan ANA, Almuktar S, Scholz M (2021) Rainfall-runoff modeling using the HEC-HMS model for the Al-Adhaim River Catchment, Northern Iraq. Hydrology, MDPI AG 8(2):58 6. Kumari N, Srivastava A, Sahoo B, Raghuwanshi NS, Bretreger D (2021) Identification of suitable hydrological models for streamflow assessment in the Kangsabati River Basin, India, by using different model selection scores. Nat Resour Res Springer 30(6):4187–4205 7. Moriasi DN, Arnold JG, van Liew MW, RL Bingner, Harmel RD, Veith TL (2007) Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans ASABE Amer Soc Agric Biological Eng (ASABE) 50(3):885–900 8. Motovilov YG, Gottschalk L, Engeland K, Rodhe A (1999) Validation of a distributed hydrological model against spatial observations. Agric Forest Meteorol 98:257–277

78

N. V. J. Reddy and R. Arunkumar

9. Sanjay Shekar NC, Vinay DC (2021) Performance of HEC-HMS and SWAT to simulate streamflow in the sub-humid tropical Hemavathi catchment. J Water Climate Change IWA Publ 12(7):3005–3017 10. Verma R, Sharif M, Husain A (2022) Application of HEC-HMS for hydrological modeling of upper Sabarmati River Basin, Gujarat, India. Model Earth Syst Environ, 1–9

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air Quality in Andhra Pradesh, India Donthi Rama Bhupal Reddy and Ramannagari Bhavani

Abstract The Novel Corona Virus Disease (COVID-19) initially observed in Wuhan city of China in 2019 induced a severe risk globally. In a drastic response to the COVID-19 pandemic, the entire country was under social and travel lockdown from 25/03/2020 to 14/04/2020 (21 days), which was extended until May 3, 2020. This lockdown had a significant influence on both the local and global economies, and it will take some time for things to return to normal. However, one major advantage of this lockdown has emerged an increase in the air quality of cities around the world. Therefore, regarding the current situation, in our study, we have analyzed the impact of lockdown on air quality in the state of Andhra Pradesh (Amravati, Rajamahendravaram, Tirupati and Vizag) and also source identification to above mentioned areas during lockdown period. The study results emphasized a statistically significant decline in pollutant concentration. Average Air Quality Index of Vizag declined more by ~34% (96.67–63.52) when compared with Amaravati by ~26% (66–48.78), Tirupati by ~31% (83.8–52.75), Rajamahendravaram by ~4% (60.95– 56.59). Future this study identified the potential sources during lockdown towards study areas with the application of HYSPLIT Back Trajectory Analysis (BTA) found that majorly from Bay of Bengal, within the state and partially from surrounding regions Odisha, Tamilnadu, Chhattisgarh, and Telangana. Keywords COVID-19 · Air quality · Back trajectory analysis

1 Introduction COVID-19 originated from Wuhan city of China in 2019 induced a severe risk globally and declared as a major disaster due to widespread global infection by the World Health Organization (WHO) [1]. According to WHO, as of May 13, 2020, ~4,369,933 COVID-19 cases have been confirmed globally including 15% deaths (https://covid19.who.int/). In case of India, first case was identified on January 30th, D. Rama Bhupal Reddy (B) · R. Bhavani JNTUACEA, Anantapur, Andhra Pradesh, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_7

79

80

D. Rama Bhupal Reddy and R. Bhavani

2020 and it’s been increased gradually. As of 2020 May 13, India crossed 75,048 COVID cases including 2440 death [2]. In a drastic response from the COVID-19 pandemic, Prime minister of India on 22 March 2020 declared a Janata curfew. After that, the entire country was under social and travel lockdown (STL) from 25/03/2020 to 14/04/2020 (21 days), which was extended until May 3, 2020, with conditional relaxation and followed by the III phase of lockdown till 17/05/2020. All businesses were forced to close as a result of social and travel restrictions, which comprises of hotels and restaurants, shops, entertainment facilities (like Malls, theatres, etc.,), transportation, etc., excluding small number of essential services like medicines, groceries, etc. Despite the lockdown, power plants continued to operate, albeit at a reduced capacity as a result of the drop in the demand for electricity from industrial and commercial units [3]. The local and global economies have already been significantly impacted by this catastrophe, and it will take some time for things to return to normal. Argumentatively, each nation’s oil usage has been significantly reduced as a result of the suspension of local transportation and routine social activities [4]. For e.g., Due to the lockdown in March 2020, resulted in a drop in fuel consumption in India by ~60–70%, which was witnessed the biggest blow in the recent 20 years [5]. Many other nations, like India, have implemented various restrictions and observed lockdowns with differing degrees of severity and implementation. The global decline in anthropogenic activity has strengthened a number of scholars to look at the environmental effects of the lockdown. Although the primary function of these restrictions was to flatten the COVID-19 infection curve, they also produced a novel experiment to examine how human activities affect air pollution. It is noticed that, there is an improvement in the air quality worldwide because of Lockdown and Travel restrictions [6, 7]. Many Countries like China, Brazil, Italy, etc., across the world reported declines in air pollution levels. In Brazil, during lockdown, there is a decline in the concentration of NO, CO, and NO2 by ~77%, ~65%, and ~54% respectively when compared with average concentration values of the same months from 2015 to 2019 reported in [8]. Muhammad et al. [4] reported there was a reduction of 20–30% in the concentration of NO2 in Italy, China, and Spain. Further, due to lockdown there is a significant decline in traffic related pollutants like Particulate matter (PM2.5, PM10), CO, NOX , and Benzene in the city of Milan [9]. A similar scenario has been noticed in the Indian context. Sharma et al. [10] reported during the lockdown reduction in concentration levels of PM2.5, PM10, CO, NO2 by 43, 31, 10, and 18% respectively by comparing with 2017–2019 pollutant concentrations to the same period. Further, [5] also reported that pollution levels in Delhi during lockdown reduced by ~41%, ~52%, ~51%, and ~28% of PM2.5, PM10, NO2 , and CO respectively, same trend has been followed in case of Kolkata Mumbai, Bangalore, and Chennai. Similar decline in the air pollution levels have been reported by [11] in India (Pune, Ahmadabad), [12] considered lockdown period (25th March to 03rd May 2020) and reported a significant decline in the concentrations of NO2 , PM2.5, and PM10 by comparing the results with 2017–2019 concentration levels over India from 134 stations. According to [12] report, every 9 out of 10 people respire polluted air that leads to an increase in premature mortality, and around 4.2 million deaths occur globally

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air …

81

which is expected to rise to 6–9 million by 2060. And very few studies have been done on South-East Indian region and most of them are not focused on the Source identification during lockdown. Therefore, considering the current situation, in our study we analyzed consequence of Social and Travel Lockdown (STL) on air quality in the state of Andhra Pradesh (Amravati, Rajamahendravaram, Tirupati, and Vizag) and also source identification to above mentioned areas during lockdown period.

2 Methodology Figure 1 represents the overview of methodological approach adopted to conduct the study. The present study assessed the impact of STL during the Novel COVID-19 break out on the air quality on major locations of Andhra Pradesh (AP) state, India as shown in Fig. 2.

Covid -19 impact assessment on air pollution and source identification in the state of Andhrapradesh during lockdown

Air Quality Assessment

Source Identification

Study areas : AndhraPradesh (Amaravathi, Rajamahendravaram, Tirupathi,Vizag

DATA COLLECTION

Real time air quality data

AIR POLLUTANTS considered in the study : PM2.5, PM10, CO, NO2,SO2

CPCB Monitoring stations Meteorological Parameters : WS,AT, RH

Time period : March, April of 2019 and From 01/01 to 24/09 of 2020 ; 24 hr average data

NOAA HYSPLIT ready GDAS DATA

7-day back trajectories at a height of 100,500 and 1000 m Trajectory shape file

MeteoInfo and TrajStat geographic information system (GIS) application

ANALYSIS

Time series analysis Lockdown variation : Before and During comparison

Yearly comparison

PSCF and CWT analysis Graphical Representation and analysis

Fig. 1 Methodological approach used in the present study

82

D. Rama Bhupal Reddy and R. Bhavani

Fig. 2 Map of study regions (Amaravati, Tirupati, Rajamahendravaram, Vizag) in Andhra Pradesh State, India along with variation in average AQI for the period before lockdown and after lockdown

2.1 Study Area and Data Source The current study assesses the variation in air quality due to STL on various locations namely Amaravati, Rajamahendravaram, Tirupati, Vizag of AP state, India. From 01/ 01/2020 to 24/03/2020 was considered as before lockdown, from 25/03/2020 to 31/ 05/2020 as during lockdown, and from 01/06/2020 to 24/09/2020 after-lockdown in

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air …

83

this study. Required air quality data of study areas for analysis was obtained from “Central Pollution Control Board (CPCB) online portal (https://app.cpcbccr.com/ ccr/#/caaqm-dashboardall/caaqm-landing)”. The 24-h average concentration data of 5 major air pollutants i.e., particulate matter with a diameter ≤ 2.5 µm (PM2.5), particulate matter with a diameter ≤ 10 µm (PM10), nitrogen dioxide (NO2 ), sulfur dioxide (SO2 ), carbon monoxide (CO) and also 3 Meteorological Parameters (MtP) Relative humidity (RH), Ambient Temperature (AT), and Wind speed (WS) were obtained from monitoring stations of each location. The above mentioned parameters continuous 24 h average data were collected for the entire study duration i.e., 01/03/ 2018 to 30/04/2018, 01/03/2019 to 30/04/2019, 01/01/2020 to 24/09/2020.

2.2 Data Analysis Trend Analysis: The following two time periods were used to examine how STL affected the air quality of study areas: Scenario 1: Lockdown period comparison was made from 01/01/2020 to 24/03/2020 (before lockdown period), 25/03/2020 to 31/ 03/2020 (During Lockdown period) 01/06/2020 to 24/09/2020 (after lockdown). Scenario 2: Yearly comparison: Air quality data obtained from CPCB for the time period of March–April 2020 was compared with data from March–April 2019. Along with these, the effect of MtP on air quality with the help of time series data was analyzed. For the entire analysis in the study, 24 h average data was considered.

2.3 HYSPLIT Backward Trajectory Analysis (BTA) The HYSPLIT (“Hybrid Single Particle Lagrangian Integrated Trajectory”) Model developed by the NOAA (The “National Oceanic and Atmospheric Administration”) Air Resources Laboratory used for calculating “air parcel trajectories, dispersion, complex transport, chemical transformation, deposition simulations by using puff or particle approaches” [13]. The model is built with a modular library structure that consists of programs for trajectories and air concentration applications. Required meteorological data fields are obtained from the forecast or meteorological models (like WRF, MM5, etc.) for calculations [14]. BTA is one of the common applications of the model, from the knowledge of paths of air mass transport from origin or source region to receptor locations that might be qualitatively measured. Back trajectories provide a mode of tracing the route of an air parcel (i.e., the origin of air masses) and develop a relationship between the source and receptor [15]. Stein et al. [16] explained model evolution, development of HYSPLIT model and also highlighted the recent applications of HYSPLIT Model. In this study, we adopted this model to identify the sources which are impacting the study areas during lockdown period. The data obtained from HYSPLIT Model is

84

D. Rama Bhupal Reddy and R. Bhavani

been visualized and analyzed with the help of Meteoinfo an ArcGIS tool [17]. Trajstat is a plugin tool used to conduct “Concentration-Weighted Trajectory” (CWT) and “Potential Source Contribution Factor” (PSCF) analysis in identifying potential sources [18]. CWT value indicates the strength of the source [19]. In CWT method, the trajectories arriving at the receptor location were weighted based on the mean concentration detected there during the trajectory’s arrival. PSCF defines “the probability that a receptor area impacted from identified regions”. Dimitriou et al. [20] used PSCF model for assessing the contribution of regional sources to urban air pollution in Athens.

3 Result and Discussion Obtained data from CPCB was analyzed in terms of daily and yearly time periods. It was observed that the concentration of pollutants varied daily and yearly due to the implementation of lockdown in the study areas (see Fig. 3). Time period of lockdown in entire India started on 24th March 2020 to 31st May 2020, and air quality data obtained from CPCB was analyzed for different time phases i.e., before-, during- and after-lockdown. Figure 3 shows the concentration of pollutants in the study region before-, during- and after-lockdown. The concentrations of all the air pollutants significantly decreased over time, according to temporal analysis of changes in their levels from the period before to during lockdown, it showed a decline in the concentration from the period before to during to after lockdown as shown in Fig. 3.

3.1 Effects of Lockdown on Air Quality When compared in the average concentration of pollutant PM2.5 before and during lockdown decrease in the concentration observed in Amaravati from 35.73 to 20.13 µg m−3 (43.66%), Tirupati from 30.66 to 18.55 µg m−3 (39.5%), Rajamahendravaram from 39.18 to 16.45 µg m−3 (58.01%), Vizag from 46.2 to 18.38 µg m−3 (60.22%). Same trend was followed for PM10 from 59.28 to 47.36 µg m−3 (20.11%), 54.67 to 40.86 µg m−3 (40.73%), 73.05 to 43.3 µg m−3 (55.74%), 100.26 to 62.27 µg m−3 (37.89%) in Amaravati, Tirupati, Rajamahendravaram, and Vizag respectively. Average Concentration of NO2 also decreased from 12.21 to 5.72 µg m−3 (53.15%) in Amaravati, 29.37 to 7.42 µg m−3 (74.73%) in Tirupati, 17.98 to 7.95 µg m−3 (55.78%) in Rajamahendravaram, and 34.36 to 25.33 µg m−3 (26.28%) in Vizag. For SO2 decrease in the concentration from 19.11 to 14.04 µg m−3 (26.53%), 5.83 to 4.78 µg m−3 (18.01%), 8.8 to 7.05 µg m−3 (19.89%), 8.45 to 8.03 µg m−3 (4.97%) in Amaravati, Tirupati, Rajamahendravaram, and Vizag respectively. It was observed that pollutant CO average concentration before and during lockdown was decreased in Amaravati from 0.58 to 0.26 mg m−3 (55.17%), Tirupati

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air … PM2.5

85

Before-Lockdown During-Lockdown

60

After-Lockdown

Concentration(μg/m3 )

50

40

30

20

10

0 AMARAVATHI

TIRUPATI

RAJAMAHENDRAVARAM

VIZAG

PM10

NO2

120 45 40 Concentration(μg/m3 )

Concentration(μg/m3 )

100

80

60

40

35 30 25 20 15 10

20 5 0

0 AMARAVATHI

TIRUPATI

RAJAMAHENDRAVARAM

AMARAVATHI

VIZAG

TIRUPATI

RAJAMAHENDRAVARAM

VIZAG

CO

SO2

1.2

25

Concentration(mg/m3 )

Concentration(μg/m3 )

1

20

15

10

5

0.8

0.6

0.4

0.2

0 AMARAVATHI

TIRUPATI

RAJAMAHENDRAVARAM

VIZAG

0 AMARAVATHI

TIRUPATI

RAJAMAHENDRAVARAM

VIZAG

Fig. 3 Impact of lockdown on air quality (before-, during-, after-lockdown)

from 0.47 to 0.44 mg m−3 (6.38%), Rajamahendravaram from 0.67 to 0.55 mg m−3 (17.91%), Vizag from 0.99 to 0.17 mg m−3 (82.82%) which is the highest % of reduction in CO concentration Percentage decline in the average concentration of CO, SO2 , NO2 , PM2.5, and PM10 in all study regions during lockdown and before lockdown is shown in Fig. 4. Highest percentage of reduction in the average concentration is observed in Vizag for CO pollutant i.e., 82.83% (see Fig. 4). Variation in the concentrations between During-Lockdown to AfterLockdown. Even after lockdown to a certain extent, it was observed that few pollutant concentrations had reduced when compared with during lockdown concentration Fig. 3. Average PM2.5 concentration after lockdown period in Amaravati and

86

D. Rama Bhupal Reddy and R. Bhavani (a) MARCH-APRIL 2019 vs. MARCH -APRIL 2020

(b) BEFORE vs. DURING LOCKDOWN' 2020 -82.83

-42.48

CO

CO

29.36

-17.9104 -6.38

-25.89

-55.17

-3.9 -21.4 0.76

SO2

11.54

AIR POLLUTANT

AIR POLLUTANT

SO2

55.27 -9

NO2 -68.51

3.8 -49.97

PM10

-29.44 -33.56 -21.19 -37.74

PM2.5

tirupathi

rajamahendravaram

amaravathi

-40

-26.53

-55.78 -74.74 -53.1532

PM10

-37.89 -40.73 -25.67 -20.11

PM2.5 -60.22 -58.01

-6.35 -12.69 -14.07 -60

-19.89 -18.01

-26.28

NO2

-19.53

-80

-4.97

vizag

-20

vizag

tirupathi

rajamahendravaram

amaravathi

0

CHANGE(%)

20

40

-39.5

60

-43.66 80

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

CHANGE(%)

Fig. 4 The % change in the concentration of pollutants a between March and April months of 2019 and 2020, b between before lockdown and after lockdown

Rajamahendravaram decreased from 20.12 to 9.99 µg m−3 and 16.45 to 12.51 µg m−3 respectively. For PM10 expect Vizag remaining Amaravati, Tirupati, and Rajamahendravaram decreased from 47.36 to 25.06 µg m−3 , 40.86 to 25.99 µg m−3 , and 43.3 to 31.99 µg m−3 respectively Fig. 3. For SO2 average concentration decreased in Amaravati and Rajamahendravaram (see Fig. 3). All the study region average pollutant concentrations showed no violation of “National Ambient Air Quality Standards (NAAQS: PM10 = 100 µg m−3 , PM2.5 = 60 µg m−3 , NO2 = 80 µg m−3 , SO2 = 80 µg m−3 based on 24 h average CO = 2 mg m−3 (based on 8 h average))” during lockdown. Variation in Pollutant Levels in March and April months of 2019 and 2020. Figure 5 shows the concentrations of PM2.5, PM10, NO2 , CO, and SO2 in the study regions for March and April months of 2019 and 2020. A significant reduction in PM2.5 concentration from 23.46 to 20.16 µg m−3 , PM10 from 64.09 to 39.91 µg m−3 , NO2 from 11.88 to 5.94 µg m−3 , CO from 0.53 to 0.51 mg m−3 , amounting to 14.06%, 37.73%, 49.97% and 3.89% reductions, respectively, whereas a rise in SO2 concentration ~55.26% (12.42–19.22 µg m−3 ) was observed in Amaravati during the months of March and April in 2019 and 2020. In case of Rajamahendravaram, the concentration of pollutants decreased from 20.40 to 17.81 µg m−3 (12.68%) for PM2.5, 52.87 to 41.67 µg m−3 (21.18%) for PM10, 0.74 to 0.54 mg m−3 (25.88%) for CO, whereas there was a rise in the concentration of NO2 from 10 to 10.38 µg m−3 (3.79%) and for SO2 from 6.53 to 7.28 µg m−3 (11.53%). A similar trend in decline of pollutant concentrations from 23.39 to 21.91 µg m−3 (6.34%) for PM2.5, 63.31 to 42.07 µg m−3 (33.55%) for PM10, 40.31 to 12.7 µg m−3 (68.51%) for NO2 , but

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air …

87

in case of SO2 and CO there is a rise from 5.85 to 5.89 mg m−3 (0.76%) and from 0.37 to 0.47 µg m−3 respectively for the location Tirupati. In case of Vizag, there is a significant decline in PM2.5 concentration from 28.92 to 23.27 µg m−3 , PM10 from 93.44 to 65.93 µg m−3 , NO2 from 32.36 to 29.44 µg m−3 , SO2 from 10.59 to 8.32 µg m−3 , CO from 0.74 to 0.43 mg m−3 , amounting to 19.53%, 29.43%, 9%, 21.4%, and 42.47% reductions, respectively was observed during month of march and April in 2019 and 2020. AQI Values in Study Areas during the Lockdown Period. AQI values were determined for March and April months of 2019 and 2020 years (see Fig. 6). As per [21] AQI is “good for the range (0–50), Satisfactory (51–100), moderately polluted (101–200), Poor (201–300) Very poor (301–400) and severe (401–500)”. In 2019 Amaravati’s AQI in the good range was 27.87% whereas it was 73.77% in 2020 and 8.2% in the moderately pollutant category in 2019 whereas zero % in 2020, which signifies a decline in the air pollutant concentrations (Fig. 6). Similarly, criteria of decline in pollutant concentration were occurred in Rajamahendravaram and Tirupati also. In Vizag there was a huge decline in the pollutant concentration in the year 2020 compared to 2019 i.e., moderately polluted category decreased from 27.6 to 6.52%. Same scenario was observed in New Delhi, Kolkata, and Mumbai, decline in AQI from 225.52, 93.8, and 65.56 in 2019 to 125.4, 65.79, and 69.79 respectively during lockdown period [22]. Mean AQI variations. Average AQI variations of study regions for the time period before and during Lockdown are presented in Fig. 2. It is clearly observed from Fig. 1 Average AQI of Vizag declined more ~34% (96.67–63.52) when compared with Amaravati ~26% (66–48.78), Tirupati ~31% (83.8–52.75), Rajamahendravaram ~4% (60.95–56.59) for Before-Lockdown to After-Lockdown period.

3.2 Meteorology and Air Quality The concentration of a pollutant during different phases of the air pollution cycle, such as pollutants’ release at the source, transport, diffusion, and their reception, is significantly influenced by meteorological factors [23]. Specifically, conditions like relative humidity (RH), wind speed (WS), and air temperature (AT) play crucial roles. In situations where RH is low and both WS and AT are high, the conditions are favorable for greater dispersion of air pollutants compared to calm weather conditions [10]. Time series analysis of all criteria air pollutants along with MtP like considered in the study for Vizag region is shown in Fig. 7. Before lockdown average values of meteorological parameters at Vizag location are AT = 31.44 °C, RH = 73.73% and WS = 1.92 m/s, During lockdown AT = 32.89 °C, RH = 71.89% and WS = 2.31 m/s. When we compare both before and during lockdown values, in case of during lockdown AT and WS are higher and RH is lower than before lockdown which are favorable for more dispersion of air pollutants. Jayamurugan et al. [24] reported the statistical correlation between RH and Particulate matter is negatively

88

D. Rama Bhupal Reddy and R. Bhavani Rajamahendravaram

40 30 20 10

Concentration ( mg m–3 )

Concentration ( μg m–3 )

10 8 6 4 2

0.7 0.6 0.5 0.4 0.3 0.2 0.1

0

0

PM2.5

0

NO2

PM10

SO2

C0 0.6

25

Concentration ( μg m–3 )

Amaravati 70 60 50 40 30 20 10

Concentration ( mg m–3 )

Concentration ( μg m–3 )

50

Concentration ( μg m–3 )

0.8

12

60

20 15 10 5 0

0

PM2.5

NO2

PM10

0.5 0.4 0.3 0.2 0.1 0

SO2

C0

70 60 50 40 30 20 10 0 PM2.5

45

Concentration ( mg m–3 )

Concentration ( μg m–3 )

Concentration ( μg m–3 )

Tirupati 40 35 30 25 20 15 10 5 0

NO2

PM10

SO2

0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

C0

80 60 40 20 0 PM2.5

PM10

35

Concentration ( μg m–3 )

100 Concentration ( μg m–3 )

Concentration ( μg m–3 )

Vizag

30 25 20 15 10 5 0 NO2

SO2

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 CO

Fig. 5 Variation in pollutants levels in 2019 and 2020 for March and April months for Rajamahendravaram, Amaravati, Tirupati, and Vizag

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air … Amaravathi_2019

Amaravathi_2020

Rajamahendravar am_ 2019

89 Rajamahendrava ram_2020 1.60 %

8.2 %

27.8 7%

26.2 3%

26.2 2%

73.7 7%

63.9 3%

Tirupathi_2019

Tirupathi_2020

73.7 8%

Vizag_2019

1.95 %

Vizag_2020

1.72 % 7.85 %

36.6 7% 63.3 3%

90.2 0%

62.3 %

36.1 %

6.52 %

27.6 %

43.4 8%

70.6 8%

82.6 %

Fig. 6 AQI comparison for the period of March–April months of 2019 and 2020

related. Sharma et al. [10] reported the average WS in North India from predominant directions (South and southwest) as 1.5 m/s and WS observed during March and April months of 2020 as ~1 m/s for southern India and ~0.7 m/s for Eastern India. Srivastava [25] investigated the relation between Meteorology, Air pollution, and COVID-19, stated that parameters like RH and AT are negatively correlated whereas solar radiation (SR) and WS are positively correlated to the coronavirus cases. And also stated there is a significant decline in the concentration of SO2 , CO, PM10, PM2.5, BC, and NOX , due to the effect of lockdown. Future [26], observed that initially there was a decline in the O3 levels due to almost stable meteorology, but in later days it increased as an increase in AT and SR which enhances the photochemical activity between nitrogen oxides and volatile organic compounds (VOC) results in ozone formation. Thus, apart from STL, (MtP) also played a significant role in reducing air pollution. Furthermore, there was a notable increase in the rainfall concentration in the state of Andhra Pradesh, from 532.9 mm (in 2019) to 694.0 mm (in 2020), specifically during the southwest monsoon period (June to September) (http://apsdps.ap.gov.in/WeatherPages/history.html).

90

D. Rama Bhupal Reddy and R. Bhavani

Fig. 7 Trend analysis representing the impact of lockdown period and meteorology on air quality in Vizag

3.3 Source Identification Even though there was travel and social lockdown imposed in the AP certain quantity of pollutant concentration been observed. To know from where the source of air mass reaching the study locations can be identified with the help of BTA. Waked et al. [27] used BTA for identifying the geographical origin of PM10 particles impacting Northern France. Similarly, [28] used BTA to know the behavior of aerosols in Spain. Limited studies done on source identification in India [29]. Mahapatra et al. [30]

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air …

91

observed that PM2.5 at receptor location are significantly impacted by sources near to study reason and long range transport of pollutants. Ravindra et al. [31] stated vehicle emissions and Crop residue burnings are predominant sources in Indo Gangetic Plain. However, all these were before COVID-19 and there were no studies done on source identification during lockdown period. The PSCF approach primarily focuses on source identification using BTA to calculate and describe potential source locations. Whereas CWT method is a concentration weighting algorithm which clearly distinguishes the source strength by assigning the concentration values at the receptor region to their corresponding trajectories. Figure 8a, c, e, g shows the Map of PSCF analysis for the study regions Amaravati, Rajamahendravaram, Tirupati, Vizag which shows the potential sources of air pollution in lockdown period. Over Amaravati, air masses are majorly from Bay of Bengal region and within the surrounding regions. Over Rajamahendravaram, major air mass reaches from Bay of Bengal and partially from neighbor states Telangana and Odisha. In lockdown period over Tirupati, air parcels are coming from south-east regions of India i.e. Tamilnadu, Bay of Bengal majorly, and a few potential sources from Karnataka and Sri Lanka. Major potential sources over Vizag location are from Bay of Bengal and the surroundings of the location, Partially from the states of Odisha and Chhattisgarh. From PSCF analysis potential source towards the study areas are majorly from Bay of Bengal and surrounding Regions. The results of the CWT analysis are shown in Fig. 8b, d, f, h creating a grid with the same resolution as the PSCF analysis. Singh and Chauhan [32] used HYSPLIT model for source identification during lockdown for the Locations New Delhi, Kolkata, Chennai, Hyderabad, and Mumbai.

4 Conclusion In response to the Novel COVID-19 outbreak, the Indian Government implemented a lockdown from March 24th, 2020, to May 31st, 2020. This measure effectively curbed the spread of the virus and led to a notable decrease in air pollution levels, however, the imposed lockdown in reaction to the COVID-19 outbreaks has affected the lives of hundreds of millions of Indians. The present study evaluated the variation in the concentration of pollutants in four major regions of Andhra Pradesh state, India. Scenario 1: On Comparison of Air quality between before lockdown to after lockdown for the study regions a significant decline in the pollution levels, Average AQI of Vizag declined more ~34% (96.67–63.52) when compared with Amaravati ~26% (66–48.78), Tirupati ~31% (83.8–52.75), Rajamahendravaram ~4% (60.95–56.59). Scenario 2: Comparison between the months of March–April 2019 and March– April 2020 showed that decline in the concentration levels of PM2.5, PM10, NO2 and CO pollutants considered in the study in all regions except CO in Tirupati, NO2 in Rajamahendravaram. Whereas, SO2 declined only in Vizag and in the rest of regions it is increased. Along with COVID lockdown, meteorological parameters also impacted the air quality.

92

D. Rama Bhupal Reddy and R. Bhavani

Fig. 8 HYSPLIT back trajectory PSCF analysis (a, c, e, g) and CWT analysis (b, d, f, h). Amaravati (a, b), Rajamahendravaram (c, d), Tirupati (e, f), and Vizag (g, h)

Comprehensive Analysis of Impact of COVID-19 Lockdown on Air …

93

In identifying the potential sources towards study areas with the application of HYSPLIT back trajectory analysis (PSCF and CWT) found that majorly within the state (local sources) and partially from surrounding regions Odisha, Tamilnadu, Chhattisgarh, Telangana. Long range transport from South-East coastal regions and Sri Lanka region. The findings would prompt the Andhra State as well as Indian government to consider ways to strictly reduce biomass burnings, and industrial and vehicular pollution in order to improve air quality and support better public health in India.

References 1. WHO (2020) Coronavirus disease (COVID-19) pandemic. https://www.who.int/emergencies/ diseases/novel-coronavirus-2019 2. MoHFW (2020) COVID-19 INDIA. https://www.mohfw.gov.in/ 3. CRISIL (2020) COVID-19 corollaries. CRISIL Limited, Mumbai, India 4. Muhammad S, Long X, Salman M (2020) COVID-19 pandemic and environmental pollution: a blessing in disguise? Sci Total Environ 138820 5. Jain S, Sharma T (2020) Social and travel lockdown impact considering coronavirus disease (COVID-19) on air quality in megacities of India: present benefits, future challenges and way forward. Aerosol Air Qual Res 20(6):1222–1236 6. Saadat S, Rawtani D, Hussain CM (2020) Environmental perspective of COVID-19. Sci Total Environ 728:138870. https://doi.org/10.1016/j.scitotenv.2020.138870 7. Bao R, Zhang A (2020) Does lockdown reduce air pollution? Evidence from 44 cities in northern China. Sci Total Environ 731:139052. https://doi.org/10.1016/j.scitotenv.2020.139052 8. Nakada LYK, Urban RC (2020) COVID-19 pandemic: impacts on the air quality during the partial lockdown in São Paulo state, Brazil. Sci Total Environ 730:139087. https://doi.org/10. 1016/j.scitotenv.2020.139087 9. Collivignarelli MC, Abbà A, Bertanza G, Pedrazzani R, Ricciardi P, Miino MC (2020) Lockdown for CoViD-2019 in Milan: what are the effects on air quality? Sci Total Environ 732 10. Sharma S, Zhang M, Gao J, Zhang H, Kota SH (2020) Effect of restricted emissions during COVID-19 on air quality in India. Sci Total Environ 728:138878 11. Yadav R, Korhale N, Anand V, Rathod A, Bano S, Shinde R, Latha R, Sahu SK, Murthy BS, Beig G (2020) COVID-19 lockdown and air quality of SAFAR-India metro cities. Urban Clim 34:100729 12. WHO (2018) World health statistics 2018: monitoring health for the SDGs. https://www.who. int/gho/publications/world_health_statistics/2018/en/ 13. Draxler RR, Hess GD (1998) An overview of the HYSPLIT_4 modelling system for trajectories. Aust Meteorol Mag 47(4):295–308 14. Draxler R, Stunder B, Rolph G, Stein A, Taylor A (2018) HYSPLIT4 user’s guide 254 15. Verma S, Venkataraman C, Boucher O, Ramachandran S (2007) Source evaluation of aerosols measured during the Indian Ocean Experiment using combined chemical transport and back trajectory modeling. J Geophys Res 112:D11210. https://doi.org/10.1029/2006JD007698 16. Stein AF, Draxler RR, Rolph GD, Stunder BJB, Cohen MD, Ngan F (2015) NOAA’s HYSPLIT atmospheric transport and dispersion modeling system. Bull Am Meteorol Soc 96:2059–2077. https://doi.org/10.1175/BAMS-D-14-00110.1 17. Wang YQ (2014) MeteoInfo: GIS software for meteorological data visualization and analysis: meteorological GIS software. Meteorol Appl 21(2):360–368. https://doi.org/10.1002/met.1345

94

D. Rama Bhupal Reddy and R. Bhavani

18. Wang YQ, Zhang XY, Draxler RR (2009) TrajStat: GIS-based software that uses various trajectory statistical analysis methods to identify potential sources from long-term air pollution measurement data. Environ Model Softw 24(8):938–939. https://doi.org/10.1016/j.env soft.2009.01.004 19. Cheng I, Zhang L, Blanchard P, Dalziel J, Tordon R (2013) Concentration weighted trajectory approach to identifying potential sources of speciated atmospheric mercury at an urban coastal site in Nova Scotia, Canada. Atmos Chem Phys 13(12):6031–6048. https://doi.org/10.5194/ acp-13-6031-2013 20. Dimitriou K, Grivas G, Liakakou E, Gerasopoulos E, Mihalopoulos N (2021) Assessing the contribution of regional sources to urban air pollution by applying 3D-PSCF modeling. Atmos Res 248:105187 21. National Air Quality Index - Report of the Expert Committee (2014) Control of Urban Pollution Series (CUPS/82/2014–15)—Central Pollution Control Board, Ministry of Environment, Forest and Climate Change, Government of India. https://app.cpcbccr.com/ccr_docs/FINALREPORT_AQI_.pdf 22. Tripathi A (2021) Air pollution in four Indian cities during the Covid-19 pandemic. Int J Environ Stud 78(4):696–717 23. Turner DB (1973) Effects of meteorological parameters on transport and diffusion. EPA. Air Pollution Training Institute Control Programs Development Division Office of Air and Water Programs 24. Jayamurugan R, Kumaravel B, Palanivelraja S, Chockalingam MP (2013) Influence of temperature, relative humidity and seasonal variability on ambient air quality in a coastal urban area. Int J Atmos Sci 2013:264046. https://doi.org/10.1155/2013/264046 25. Srivastava A (2021) COVID-19 and air pollution and meteorology-an intricate relationship: a review. Chemosphere 263:128297 26. Garg A, Kumar A, Gupta NC (2021) Comprehensive study on impact assessment of lockdown on overall ambient air quality amid COVID-19 in Delhi and its NCR, India. J Hazard Mater Lett 2:100010 27. Waked A, Bourin A, Michoud V, Perdrix E, Alleman LY, Sauvage S, Delaunay T, Vermeesch S, Petit J-E, Riffault V (2018) Investigation of the geographical origins of PM10 based on long, medium and short-range air mass back-trajectories impacting Northern France during the period 2009–2013. Atmos Environ 193:143–152. https://doi.org/10.1016/j.atmosenv.2018. 08.015 28. Piñero-García F, Ferro-García MA, Chham E, Cobos-Díaz M, González-Rodelas P (2015) A cluster analysis of back trajectories to study the behaviour of radioactive aerosols in the south-east of Spain. J Environ Radioact 147:142–152. https://doi.org/10.1016/j.jenvrad.2015. 05.029 29. Banerjee T, Murari V, Kumar M, Raju MP (2015) Source apportionment of airborne particulates through receptor modeling: Indian scenario. Atmos Res 164:167–187. https://doi.org/10.1016/ j.atmosres.2015.04.017 30. Mahapatra PS, Sinha PR, Boopathy R, Das T, Mohanty S, Sahu SC, Gurjar BR (2018) Seasonal progression of atmospheric particulate matter over an urban coastal region in peninsular India: role of local meteorology and long-range transport. Atmos Res 199:145–158. https://doi.org/ 10.1016/j.atmosres.2017.09.001 31. Ravindra K, Singh T, Mor S, Singh V, Mandal TK, Bhatti MS, Gahlawat SK, Dhankhar R, Mor S, Beig G (2019) Real-time monitoring of air pollutants in seven cities of North India during crop residue burning and their relationship with meteorology and transboundary movement of air. Sci Total Environ 690:717–729. https://doi.org/10.1016/j.envpol.2017.08.016 32. Singh RP, Chauhan A (2020) Impact of lockdown on air quality in India during COVID-19 pandemic. Air Qual Atmos Health 13(8):921–928

Development of Mobile Application for Assessing Urban Heat Island (UHI) Using Geospatial Techniques a Case Study of Chennai City S. Jayalakshmi

Abstract Urban Heat Island (UHI) is the phenomenon where urbanization results in an increase in surface temperature among different locations within the city. UHI hotspots not only lead to poor air quality and make people’s health at higher risk, but they also tend to magnify the heat stress and level of thermal discomfort experienced by the people. This study aims to find the UHI spots using thermal remote sensing based on satellites, for the estimation of surface temperature, over a continuous spatial and temporal scale and to develop a mobile application indicating the spatial pattern of UHI and heat stress. Wet Bulb Globe Temperature (WBGT) data collected at various locations across Chennai city was evaluated to obtain the indices reflecting risk levels of heat stress in each area. This was subsequently analyzed in a GIS environment, along with the disaggregated Land Surface Temperature (LST) data, to arrive at valuable information that was used to delineate the hotspots of high heat stress and UHI intensity in the city. Finally, this data was exported to a mobile platform (Android) and an application indicating the spatial pattern of UHI and heat stress was developed, which shows the heat risk zones, mitigation measures, etc. This study confirmed the existence of UHI effect in Chennai city during summer. Temperature difference was found to be even as high as 6–7 °C in many parts of the city. The intensity of UHI was established to be strongly dependent on urban factors such as the density of built-up areas, vegetation cover and presence of water bodies. It was shown that such adverse heat conditions deteriorated the urban environment causing health problems. The results of this study indicate that the highest thermal stress is found in the South-Western and Northern part of the city, which is predominantly crowded, constructed (built-up), industrial and commercial areas. Keywords Urban heat island · Thermal data · Wet bulb globe temperature · Land surface temperature · Mobile application

S. Jayalakshmi (B) Institute of Remote Sensing, College of Engineering Guindy, Anna University, Chennai, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_8

95

96

S. Jayalakshmi

1 Introduction As urban regions develop, modifications occur in their landscape. Man-made features like roads, buildings, and other infrastructure replace natural land cover such as vacant land and vegetation. Areas that were once permeable and moist become impermeable and dry [1]. These changes result in urban areas becoming warmer than their rural surroundings. This leads to the formation of an “island” of higher temperatures in the urban landscape, termed as a surface ‘Urban Heat Island’ (UHI). Surface temperature indicates the heat energy given off by features such as vacant land, buildings, and other surfaces. Thermal instruments mounted on satellites measure this surface temperature of features on the ground, and provide better geographic coverage than ground-based instruments which are normally used for recording air temperatures [2]. There are some drawbacks in satellite data also. A combined use of satellite data for surface temperatures and data from monitoring stations for air temperatures offers the most complete picture of a city’s heat island [3]. Due to industrialization, urban landscapes are often characterized by the substitution of natural surfaces like vegetation by man-made surfaces such as high albedo parking spaces, concrete areas, asphalt roadways etc. Such an urbanization process has been found to affect the thermal environment in cities [4]. The magnitude of UHI intensity is generally found to be greater at night than during the day [5]. The main cause of this phenomenon is the higher rate of nocturnal cooling of the vacant areas around cities when compared with that of densely built-up localities [6]. Satellitebased thermal remote sensing data can provide temperature information of different land use categories at regional levels. This can be utilized to identify, map and study the spatial distribution of the surface temperature of UHI hot spots. Voogt and Oke [7] stated that atmospheric UHI is observed from ground-based air temperature measurements taken from weather stations, whereas surface UHI is observed from thermal remote sensors that record the upwelling thermal radiance emitted by the surface area that lies within the instantaneous field of view of the sensor. Schwarz et al. [8] found that the eleven indicators that were analyzed individually reveal diurnal and seasonal patterns, but show rather low correlations over time and for single points in time, the different indicators show only weak correlations, although they are supposed to quantify the same phenomenon. UHIs not only have fundamental impacts on urban climate, air quality and economics, but also on the public health [9]. Steeneveld et al. [10] applied an approximated ‘wet bulb globe temperature’ (AWBGT) index, that is based on the measurement of basic environmental variables, to quantify the measure of human comfort. While the scientific community has embraced the use of satellite imagery as a tool for phenological studies, BeeBox, a web-based application developed by Arundel et al. [11] attempts to make this same information available to a more general audience (beekeepers). Finer spatial resolution thermal data comes with the drawback of poorer temporal resolution was used in most of the studies [12–15]. This study focuses on mapping the spatial distribution of UHI and heat stress levels in the metropolitan city of

Development of Mobile Application for Assessing Urban Heat Island …

97

Chennai, and analyzing the correspondence between them. It incorporates an automated approach to downscale satellite-based surface temperature data, for delineation of heat island hotspots. This provides the basis for subsequent development of a useful mobile application featuring heat stress in various heat zones, and supplying vital precautionary information to alleviate the adverse impacts of heat stress on human health.

1.1 Study Area Chennai is one of the four metropolitan cities in India. It is the capital of Tamil Nadu state which is also a very famous south Indian eastern coastal city located in North-Eastern part of Tamil Nadu. Its geographical extent is 12° 51' 3'' N to 13° 18' 2'' N latitude and 80° 00' 00'' E to 80° 14' 15'' E longitude (Fig. 1). The city is divided as North, Central, South and West. North Chennai and Central Chennai are industrial and commercial centers of the city respectively. The known residential places of South and West Chennai are now turned into commercial hubs at rapid manner. The city is growing dramatically with diversified economic base along with industrial and commercial development. The city is one of the known automobile assembly and production centers in the world and also it is becoming an IT center and back office. Due to the dramatic increase in economic growth of the city, the built-up area was increased exponentially which resulted in the exhaustive change in land use/land cover. It is observed that the average population density was also arbitrarily increased in the city for the past few years which is the cause for the increased concrete structures. The climatic and weather condition of Chennai is determined by its geographical location. The city’s weather and climate is reliable and has mild variation in the seasonal temperature due to its close proximity to the sea. Though Chennai comes across three seasonal variations such as summer, monsoon and winter, mostly it experiences hot and humid weather conditions only. The city experiences the maximum temperature variations from 38 to 42 °C during April to June i.e. summer season.

2 Data Sets Used 2.1 Satellite Datasets The methodology consists of remotely sensed satellite thermal data analysis using data from the OLI/TIRS and MODIS sensors on-board NASA’s Landsat-8 and Terra satellites respectively. Surface reflectance (MOD09GA) data corrected for atmospheric errors and LST data products (MOD11A1) of MODIS for Chennai (H25V07

98

S. Jayalakshmi

Fig. 1 Location of Chennai

MODIS tile). The images are in HDF-EOS format, in sinusoidal projection. Resampled LST data at sampling intervals of 960 and 240 m for surface reflectance products using geometrically registered UTM projection Zone 44N, WGS 1984 Datum of MODIS Reprojection Tool; thus, the pixel dimensions are ensured to be in multiples of 30 m. The output images are to be stored in GeoTIFF format. Concurrent Landsat8 images for Chennai (Path 142, Row 51), with the same geographic projection as MODIS, are resampled to 60 m, by changing ‘Cell Size’ in ‘Reproject’ option in ERDAS Imagine software (for NL-disTrad algorithm). The datasets are acquired for the date 12 May 2017, for which both sensor data are available.

2.2 Other Datasets Administrative ward boundary of Chennai City is used to identify urban and rural areas. It is procured in shapefile format, in UTM Zone 44N (WGS84) co-ordinate system. Field measurements to estimate WBGT were carried out using WBGT monitoring devices across selected locations in the city during summer season (May 2017). It recorded certain essential parameters such as temperature, relative humidity, etc. A perception social survey was also conducted, to ascertain self-related symptoms of heat-related illnesses. It was carried out among individuals residing in different parts of the urban area of Chennai.

Development of Mobile Application for Assessing Urban Heat Island …

99

3 Methodology The overall methodology of the study is depicted in Fig. 2 which encompasses two components of work. The first component is concerned with mapping of UHI in the study area, using MODIS LST that has been disaggregated to Landsat-8 scale. The second component involves mapping the heat stress index (WBGT) and heat stress-related illnesses, in the study area. Finally, the development of a mobile application using the results of the overlay analysis, to issue alerts and suggest mitigation measures.

3.1 UHI Mapping The fine-resolution temperature map was prepared using LST dataset obtained by manual processing of Landsat-8 thermal data product. The urban and rural area temperature difference was selected as the indicator to assess UHI intensity. To identify urban/rural areas in the temperature data layer, the raster was overlaid with

Fig. 2 Overall methodology

100

S. Jayalakshmi

an administrative boundary map of Chennai. UHI intensity indicator was computed using Eq. (1): UHIintensity = LSTurban − LSTrural

(1)

where: LSTurban LST of a urban pixel. LSTrural LST of a rural pixel (taken as 25 °C). UHIintensity UHI intensity. Temperature was classified in four different classes such as 32 °C for hotspot identification. The mean LST within each ward was determined using Zonal Statistics tool in QGIS.

3.2 Heat Stress Index (WBGT) Mapping For human beings, the experience of thermal discomfort is related to the UHI effect [10]. Wet Bulb Globe Temperature (WBGT) was used to assess the heat stress. The WBGT combines the effect of the four main thermal components affecting heat stress: air temperature, humidity, air velocity and radiation, as measured by the dry bulb, wet bulb and globe temperatures [16]. The index can be estimated from the Wet Bulb Globe Temperature using Eq. (2) [17]: WBGT = 0.7Tw + 0.2Tg + 0.1Ta

(2)

where: T w Natural wet bulb temperature. T g Black globe temperature. T a Air temperature (dry bulb temperature). A WBGT monitoring device was used to assess WBGT accurately in several locations across Chennai. Model QuesTemp°34 heat stress monitor with dry bulb temperature accuracy of ±0.5 °C for the temperature range of 0–120 °C and ±5% accuracy for relative humidity (RH) in the range between 20 and 95% RH manufactured by Quest Technologies, USA was used.

3.3 Heat Stress-Related Illnesses Mapping Assessments of heat vulnerability are done based on self-rated health questions in the health sciences and other fields because, over decades of research, they have been found to be strong correlates of clinical conditions [18]. A perception–social

Development of Mobile Application for Assessing Urban Heat Island …

101

survey was conducted across various localities in Chennai, to study the interactions between humans and environment. Opinions of the respondents about the changes in quality of life in their neighborhoods of the natural environment and in their location are collected by asking questions emphasizing climate and land use. Questions on experiencing heat-related symptoms and self-reported heat illnesses were also included.

3.4 Development of Mobile Application With the different maps that were obtained in the study, a UHI—heat stress mobile application was developed, using which a user can see if they might be in a heat stress prone area. It can serve as a visual indicator of the heat stress index and ‘risk level’ specific to their current geographical location. The application was developed on the Android platform using Android Studio (Version 3.0.1) software. Google Maps Android API, was used to include Google Maps and add customized mapping information (LST, WBGT, Health Impact, NDVI, NDBI, NDWI, Landuse Proportion) in the application. This was done by embedding maps into an activity as a fragment with a simple XML snippet. The APK file can be installed and run on any Android device.

4 Results and Discussion 4.1 Spatial Distribution of UHI Hotspots UHIs were identified (Fig. 3) from the thermal map, by the difference between the temperature of each pixel with that of a rural pixel (taken as 25 °C). This was accomplished using Raster Calculator tool in QGIS. Temperature difference was found to be even as high as 6–7 °C in many parts of the city. The UHI hotspots were demarcated (Fig. 4a) by categorizing the temperature classes and highlighting the higher value classes. The places which were known to have congested built-up concrete areas/crowded urban areas and also, poor vegetation cover, such as Old Washermanpet (George Town), Choolaimedu, Virugambakkam, were observed to be within the UHI hotspots. Temperatures were cooler than these areas, in central and southern regions of the city, where urbanization process was recognized to be more recent, and hence were of less building density. Both of these zones were known to be of less built-up concrete area. The occurrence of green cover was also known to be higher (such as Guindy), or proximity to water bodies was known to be closer (such as Velachery, Neelankarai). Dense greenery showed its positive impact to the thermal environment,

102

Fig. 3 Urban–rural temperature difference indicator: UHI

Fig. 4 Temperature maps: a UHI hotspots b ward-wise LST mean

S. Jayalakshmi

Development of Mobile Application for Assessing Urban Heat Island …

103

as these places were outside the UHI hotspots. The mean LST within each ward was computed using Zonal Statistics tool in QGIS (Fig. 4b).

4.2 Spatial Distribution of WBGT World’s widely used heat stress index of WBGT index was used in this study. It combines the influence of four environmental parameters into a single number. Measurements were taken at different categories of localities to cover all weather conditions, such as near park/dense trees (Guindy, AC Tech, Viyasarpadi), close to water bodies (Kottur Garden, Velachery, Ramapuram), crowdly constructed area (George town, Cholaimedu, Virugambakkam), near hilly region (St. Thomas mount), near sea shore (Triplicane, Besant nagar, Neelankarai). Simultaneous measurements of natural wet temperature, air temperature, air velocity, relative humidity and globe temperature were taken for five days at halfhourly intervals, at selected locations in the city, using the instrument ‘Questemp 34’. The point measurements at a site were generalized for the corresponding entire ward in GIS. In this way, point information was converted to regional information. Figure 5 graphically represents the results of summer heat stress on a map, prepared using QGIS.

Fig. 5 WBGT map

104

S. Jayalakshmi

Fig. 6 Results of health impact study

According to ISO-7243, the temperature ranges of 33 °C, 33–35 °C and more than 35 °C were classified as safe, caution and stress respectively. In summer, most parts of Chennai measured higher value of WBGT than the reference value of 33 °C. Only very few safe conditions were found throughout the city, in the wards where WBGT was studied. This situation was characterized by a warning status of heat stress.

4.3 Examination of Health Impact Based on the questionnaire perception survey conducted across the various localities, it was determined that a majority of the working class people in the city felt that heat did induce some health problems and caused an adverse effect on their overall wellbeing. The results of the health impact of heat stress have been represented as a graphical bar chart in Fig. 6.

4.4 Interpretation from GIS Analysis A land use/land cover (LULC) map of Chennai (Fig. 7) was prepared using Landsat-8 OLI bands 5,4,3. The projection and extent of the resulting raster were changed to be the same as that of the other rasters (such as LST). Satellite based temperature measurement is found to be a potential way to monitor the temperature in a highly human influenced urban environment. The thermal UHI hotspot map which indicated that northern and central western areas were the hottest of the city, was overlaid with the WBGT map (Fig. 8). It is observed from the LULC map (Fig. 7) that the regions which were known to have dense built-up areas and subsequently, poor vegetation cover, fell within the

Development of Mobile Application for Assessing Urban Heat Island …

Fig. 7 LULC map—Chennai

Fig. 8 Overlay of UHI hotspot map and WBGT map

105

106

S. Jayalakshmi

UHI hotspots and had higher WBGT values. The reverse case: lower temperatures, outside the UHI hotspots, and lower WBGT values were found in recently urbanized wards, where vegetation cover was higher and also most of them were closer to inland water bodies or the coast. The results of this study, indicate that the highest thermal stress is found in the South-Western and Northern parts of the city, which are predominantly crowded, constructed (built-up), industrial and commercial areas, such as George Town (Old Washermanpet), Virugambakkam, etc. The map also depicts that in the southern part of the city and to some extent, in its central parts corresponding to green regions (park areas, hilly areas, etc.) or blue regions (areas closer to water bodies, coastal areas, etc.), respectively, heat stress condition is relatively lesser, as the WBGT values are lower. The significant relationship between UHI hotspots, WBGT, health impacts and land use/land cover, especially in summer shows that some specific localities of the city suffer the most heat stress. Addressing heat stress and UHI within continuous space has thus been explored to be an effective approach for better planning and management of the adverse effects of heat that can be adopted by the authorities.

4.5 Information from Mobile Application The rich maps provided by Google, when augmented with custom markers and custom geographic data (such as LST, WBGT, etc.), allow the users to explore the area of interest in much depth and obtain valuable, easy-to-read information. Thus, in order to present the results of this UHI-WBGT study to an end-user without GIS knowledge, the Google Maps Android API was used to develop a simple mobile application. If the GPS is switched on, the user’s current location is detected by the app, and a marker is placed on the Google Maps screen. Google Maps is designed to zoom into the study area (Chennai), with a ward boundary map overlaid on it. The map is color coded to indicate the 3 levels of heat stress within the city—high, moderate and low, based on the results of GIS analyses, on LST, WBGT and Health Impact. The display of the app icon and home screen is shown in Fig. 9.

5 Conclusion In this study, satellite-derived temperature data was found to be suitable for studying the spatial pattern of UHI intensity distribution. The implications of UHIs on human heat stress were also addressed from the perspective of continuous space. From this study, it is proved that Chennai is affected by UHI during summer. South-Western and Northern part of the city which have dense buildings shows UHI. The developed mobile app will give information about UHI affected areas and mitigation measures to the users. Surface temperature measured using satellite-based instruments, can be

Development of Mobile Application for Assessing Urban Heat Island …

107

Fig. 9 Launching the app: a app icon b home screen

utilized to estimate air temperature indirectly, this estimates are less reliable than direct ground-based measurements. WBGT can’t provide an accurate assessment of heat stress; it will give only a general guideline. More accurate information about heat stress can be obtained by measuring the specific elements of the thermal environment.

References 1. Alfraihat R, Mulugeta G, Gala TS (2016) Ecological evaluation of urban heat island in Chicago City, USA. J Atmos Pollut 4(1):23–29 2. Meng F, Liu M (2013) Remote-sensing image-based analysis of the patterns of urban heat islands in rapidly urbanizing Jinan, China. Int J Remote Sens 34(24):8838–8853 3. Dousset B, Gourmelon F, Laaidi K, Zeghnoun A, Giraudet E, Bretin P, Mauri E, Vandentorren S (2010) Satellite monitoring of summer heat waves in the Paris metropolitan area. Int J Climatol 31(2):313–323 4. Roth M, Oke TR, Emery WJ (1989) Satellite-derived urban heat islands from three coastal cities and the utilization of such data in urban climatology. Int J Remote Sens 10(11):1699–1720 5. Oke TR (1982) The energetic basis of the urban heat island. Q J R Meteorol Soc 108:1–24 6. Nichol JE, Hang TP (2012) Temporal characteristics of thermal satellite images for urban heat stress and heat island mapping. ISPRS J Photogramm Remote Sens 74:153–162 7. Voogt JA, Oke TR (2003) Thermal remote sensing of urban climates. Remote Sens Environ 86:370–384 8. Schwarz N, Lautenbach S, Seppelt R (2011) Exploring indicators for quantifying surface urban heat islands of European cities with MODIS land surface temperatures. Remote Sens Environ 115:3175–3186 9. Beniston M (2004) The 2003 heat wave in Europe: a shape of things to come? An analysis based on Swiss climatological data and model simulations. Geophys Res Lett 31:2022–2026 10. Steeneveld GJ, Koopmans S, Heusinkveld BG, Hove LWA, Holtslag M (2011) Quantifying urban heat island effects and human comfort for cities of variable size and urban morphology in the Netherlands. J Geophys Res 116(D20129) 11. Arundel J, Winter S, Gui G, Keatley M (2016) A web-based application for beekeepers to visualise patterns of growth in floral resources using MODIS data. Environ Model Softw 83:116–125

108

S. Jayalakshmi

12. Chapman S, Thatcher M, Salazar A, Watson JEM, McAlpine CA (2019) The impact of climate change and urban growth on urban climate and heat stress in a subtropical city. Int J Climatol 39:3013–3030 13. Keikhosravi Q (2019) The effect of heat waves on the intensification of the heat island of Iran’s metropolises (Tehran, Mashhad, Tabriz, Ahvaz). Urban Clim 28:100453 14. Masek JG, Wulder MA, Markham B, McCorkel J, Crawford CJ, Storey J, Jenstrom DT (2020) Landsat 9: empowering open science and applications through continuity. Remote Sens Environ 248:111968 15. Jabbar HK, Hamoodi MN, Al-Hameedawi AN (2023) Urban heat islands: a review of contributing factors, effects and data. IOP Conf Ser: Earth Environ Sci 1129:012038 16. Song Y, Wu C (2017) Examining human heat stress with remote sensing technology. GIScience & Remote Sens 54 17. Chinnadurai J, Venugopal V, Kumaravel P, Paramesh R (2016) Influence of occupational heat stress on labour productivity—a case study from Chennai India. Int J Prod Perform Manag 65(2):245–255 18. Belanger D, Gosselin P, Valois P, Abdous B (2015) Neighbourhood and dwelling characteristics associated with the self-reported adverse health effects of heat in most deprived urban areas: a cross-sectional study in 9 cities. Health Place 32:8–18

Drones as an Alternate Communication System During Calamities D. S. Vohra, Pradeep Kumar Garg, and Sanjay Kumar Ghosh

Abstract In this age of rapidly advancing technology, numerous concepts have captured the attention of researchers, and drone technology is no exception. Researchers are significantly captivated by the myriad uses of drones, encompassing civil applications such as the scrutiny of power infrastructure, surveillance of wildlife, transportation of medical supplies to remote locales, identification of forest fire outbreaks, and assessment of landslides. Additionally, they are intrigued by the military potentials, including real-time monitoring, surveillance operations, patrolling activities, and demining efforts. Despite the wide range of applications already explored, some countries still have untapped drone potential. One such area is the utilization of drones as communication relays during natural disasters when conventional communication lines are disrupted. This approach could prove highly beneficial in rescuing affected individuals, as the aerial node created by drones would enable people to communicate with rescue teams using mobile phones or ordinary landline telephones, even when devastating natural calamities like tsunamis, earthquakes, or floods have destroyed traditional communication towers. Keywords Communication relay · Path loss model · Signal-to-Noise Ratio (SNR)

D. S. Vohra (B) · P. K. Garg · S. K. Ghosh Indian Institute of Technology (IIT) Roorkee, Roorkee, India e-mail: [email protected] P. K. Garg e-mail: [email protected] S. K. Ghosh e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_9

109

110

D. S. Vohra et al.

1 Introduction While most people commonly associate drones with aerial photography and videography, only a few have considered their potential as communication relays. Particularly during natural disasters like tsunamis, floods, earthquakes, and landslides, traditional communication lines often get destroyed, leaving the affected areas without reliable communication. In our research, we propose using drones as relays to bridge the communication gap between ground devices and terrestrial base stations far away from the affected region. The idea is to strategically position the drone in the air, allowing it to act as an intermediary link, reconnecting the ground communication devices to the terrestrial network until regular communication facilities are fully restored. This approach could offer a crucial lifeline to maintain communication during times of crisis and help coordinate rescue and relief efforts efficiently [7, 9, 19].

2 Preceding Work Over time, many researchers have delved into diverse strategies for using drones as communication relays. These strategies frequently center around two primary facets: the mitigation of transmit power to economize energy and the enhancement of quality of service (QoS) to ensure proficient communication [18, 25]. Several scholars have grounded their investigations in path loss models to scrutinize the viability of employing drone communication relays. Furthermore, specific scholars have directed their focus toward augmenting drones’ energy efficiency, aiming to refine their efficacy as communication relays. They have undertaken inquiries into the deployment of drones to enhance the dependability of their services. To systematically categorize the wide array of deployment techniques and roles that drones assume as communication relays, Fig. 1 presents a classification scheme grounded in diverse criteria. Drone’s deployment as a relay node based on different functions.

Transmission power reduction Backhaul communication delay Maximizing QoS Increasing throughput

Fig. 1 Deployment classification of the drone as a relay [22]

Drones as an Alternate Communication System During Calamities

111

3 Mathematical Model 3.1 Logical Settings for the System To optimize the system settings for effective communication, we begin by considering a total of M wireless devices that are positioned far from the central communication base station, and due to their limited power capacity, they cannot establish direct communication with the remote base station [5]. To address this issue, we aim to strategically position a drone at coordinates (X u , Y u , Z u ) to serve as a relay for the M wireless devices, enabling them to establish communication with the remote base station [4]. Figure 2 illustrates this concept pictorially. To ensure successful communication, we propose using Frequency Division Multiple Access (FDMA) as the multiple access technique for drones. FDMA allows the drone to transmit information on a specific sub-channel, ensuring that the Signal-to-Noise Ratio (SNR) for communication is greater than or equal to a predefined minimum threshold value (SNRth ). This approach minimizes interference and facilitates reliable communication for each user with one dedicated sub-channel per drone [23]. Furthermore, we assume the maximum communication power the aerial drone can transmit is denoted as pmax . These system settings and techniques will enable efficient communication between the M wireless devices and the remote base station, leveraging the drone as a reliable relay.

A Base station to drone

V G P L (dB)

Drone to ground user

140 120 100 80 60 40 20 0 0.5 to 1

1 to 1.5

1.5 to 2

Horizontal Distance --------------------> (x104) Fig. 2 Path loss model equilibrium [22]

2 to 2.5

112

D. S. Vohra et al.

3.2 Investigation of Path Deficit Models For a comprehensive assessment of path loss models, we initiated the computation of these communication models connecting the primary remote communication station with the drone and subsequently between the drone, serving as a relay, and the terrestrial users. Moreover, determining the path loss model is also feasible between the drone and the ground-based user [20, 27]. Equation 1 embodies the path loss model that links the base station and an airborne relay node. This equation serves as a mathematical representation that encapsulates path loss, taking into cognizance variables like distance, frequency, and other pertinent parameters between the base station and the airborne relay drone. The pertinence of this model is underscored by its pivotal role in the comprehension and enhancement of communication performance within the drone relay system [22]. L BS−UAV (d2D , θ ) = 10α log(d2D ) + A(θ − θo ) exp(−(θ − θo )/B) + ηo + N (0, aθ + σo )

(1)

where

] √[ (X u − X BS )2 + (Yu − YBS )2 represents the flat-plane separation between a base remote communication station and a drone, θ Receding Angle, α Path Attenuation Example, A Amplifier for Substantial Path Attenuation, Angular Deviation, θo Angle Multiplier, B ηo Compensation for Significant Path Attenuation, N (0, aθ + σ o ) is a Random variable (Gaussian) in which ‘a’ = shadowing of drone slope, σ o = offset for shadowing of drone. d2D

Equation 2 delineates the path loss model connecting drones and ground users (denoted as i ∈ M), where the variable i represents individual ground users, with the total count not exceeding M [22]. LUAV−i (d3D , φ) = P(LOS) · L LOS + P(NLOS) · L NLOS where

(2)

] ( √[ (X u − Xi )2 + Y u − Y i 2 + Z u 2 denotes the spatial separation in three dimensions between a drone and a ground user, P (LOS) Ensuring the likelihood of a clear line of sight (LOS) at a specific elevation angle φ, P (NLOS) Ensuring the non-likelihood of a clear line of sight (LOS) at angle φ, L LOS Mean path attenuation for line of sight (LOS) trajectories, L NLOS Average path attenuation for non-line of sight (NLOS) scenarios. d3D

Drones as an Alternate Communication System During Calamities

113

3.3 Path Loss Model Equilibrium Attaining an optimal equilibrium between the distances spanning from the base remote communication station to the drone and from the drone to the ground user stands as a pivotal endeavor, bearing profound significance in harnessing path loss and enacting the most influential power allocation mechanism [10, 17]. This equilibrium is paramount in securing a seamless and dependable communication environment within a drone relay system. As the distance between the drone and the ground user dwindles, the path loss originating from the base station to the drone accentuates [11]. Conversely, when the gap between the drone and the ground user broadens, the path loss between the base station and the drone abates. To maximize advantages for ground users, identifying a strategic vantage point for the drone becomes imperative, where the collective impacts of path losses from the base station to the drone and from the drone to the ground users are minimized. Once attained, this equilibrium empowers the drone to function proficiently as a relay, adeptly catering to an expanded spectrum of ground users, extending the scope of communication, and ushering in enhanced coverage. The meticulous selection of an optimum drone placement catalyzes the overall communication performance, elevates connectivity, and bolsters reliability for users in locales marked by limited communication infrastructure. Consequently, a methodical contemplation of path loss and judicious power allocation emerges as a linchpin in the triumphant deployment of drone communication relays across diverse applications (Fig. 3).

4 Devising of Mathematical Problem Initially, the focus was on triple-play communication services, encompassing data, voice, and video transmission between a geographically distant communication base station and a terrestrial user. This intricate communication process is made possible through the intermediary role of a drone functioning as a relay. The quantification of the communication link’s capability between the distant communication base station and the drone is expressed by Eq. 3. Equation 3 quantifies the data rate or throughput of the communication link between the base station and the drone. The capacity of this link plays a critical role in determining the overall performance of the communication system. It depends on various factors, including the available bandwidth, signal-to-noise ratio, modulation scheme, coding efficiency, and other

Fig. 3 Framework configuration [22]

114

D. S. Vohra et al.

channel characteristics. By optimizing the link’s capacity, we can enhance data transmission efficiency between the base station and the drone, thus ensuring seamless communication for triple-play services to the ground users. Equation 3 and the parameters involved are necessary to understand the communication link’s capacity and its implications for the triple-play services [22]. CBS−UAV = BBS · log2 (1 + SNRBS−UAV )

(3)

where BBS Bandwidth emitted by the ground base station, SNRBS−UAV Drone’s SNR. The data capacity is [22] CUAV−i = BUAV−i · log2 (1 + SNRUAV−i )

(4)

where BUAV−i Drone’s originating bandwidth, SNRUAV−i Ground user’s SNR. The ascribed bandwidth ceiling for a ground user is estimated to stand at BUAV/ M, with BUAV signifying the drone’s bandwidth and M denoting the upper limit of wireless devices [16, 21]. Subsequently, the requisite data rate Ri for each wireless device is expounded in Eq. 5 [22]. ( pi = 2 Ri

M/BUAV−1

)

N · L UAV−i

(5)

The succeeding lines elucidate the meanings of the assorted variables in Eq. 5. L UAV−i Average attenuation experienced between the drone and the ground user i. Power via noise. N The objective is to strategically position drones to guarantee optimal communication for the maximum number of affected ground users [2, 3, 13]. This optimization challenge is thus outlined through Eqs. 6 and 7 [22]. Maximize

Σ

(i = 1 to M) BUAV−i · log2 (1 + SNRUAV−i )

(xu , yu , z u , P) subject to

Σ

(6)

(i = 1 to M)BUAV−i · log2 (1 + SNRUAV−i ) ≤ BBS · log2 (1 + SNRBS−UAV )

SNRUAV−i ≥ SNRth , ∀i ∈ M, Σ (i = 1 to M) pi ≤ p max , pi ≥ 0, ∀i ∈ M xmin ≤ X u ≤ xmax , ymin ≤ Yu ≤ ymax , Z min ≤ Z u ≤ Z max

(7)

Drones as an Alternate Communication System During Calamities

115

4.1 Constraints for Mathematical Problem The constraints described in the scenario play a crucial role in determining the optimal placement and operation of the drone as a communication relay. These constraints are elaborated in brief in succeeding lines. 1. Capacity Constraint. The communication link’s capacity established between the remote communication base station and the drone ought to adhere to a principle wherein it remains either lower than or equal to the capacity inherent in the communication link uniting the drone with the ground user. 2. Signal-to-Noise Ratio (SNR) Constraint. The SNR of the ground user’s communication should always be greater than or equal to the threshold value SNRth . This constraint ensures that the received signal quality is sufficient for reliable communication. 3. Power Constraint. The total power consumed by the drone should not exceed the maximum emanating level pmax . This constraint limits the drone’s power usage to ensure energy efficiency and avoid overwhelming the system. 4. Non-Negative Power Constraint. The power consumed by the drone should always be greater than or equal to zero, indicating that the drone cannot consume negative power. The constraints above cited earlier collectively serve as guiding principles that steer the decision-making process for the placement of the drone and the allocation of power. This strategic decision-making aims to maximize communication efficiency while staying within the bounds of physical and operational limitations. This scenario is visually depicted in Fig. 4, showcasing the intricate interplay involving the remote communication base station, the drone, and the ground users. The

Fig. 4 Scenario description of remotely placed base station, drone, and ground users [22]

116

D. S. Vohra et al.

illustration underscores the drone’s role as a relay, orchestrating the transmission of signals between the base station and the ground users, all while diligently considering the established constraints. The overarching goal of optimizing both the placement and operation of the drone, in alignment with these constraints, holds paramount importance in ensuring an effective and reliable communication framework. This optimization is pivotal in facilitating the successful delivery of triple-play services to ground users [26].

5 Methodology The Particle Swarm Optimization (PSO) algorithm is widely employed for tackling optimization challenges, encompassing domains like drone placement. Its essence lies in iteratively adjusting particle velocities and positions to determine optimal configurations or positions. Within the context of this research, the drone was statically positioned at a fixed altitude, and the analysis revolved around the throughput of users impacted by disasters. Various power allocation algorithms were explored, including equal power distribution, water-filling, and modified water-filling algorithms. The outcomes disclosed that the water-filling algorithm consistently outperformed its counterparts, even the modified variant, across diverse distances between ground users and drones. Notably, this superiority persisted beyond 3000 m, exemplifying the robustness and effectiveness of the water-filling algorithm in heightening user throughput. These assertions are corroborated by Fig. 5, which consistently underscores the supremacy of the water-filling algorithm, regardless of the distance between ground users and drones. This research effectively

Modified Water Filling vs Water Filling vs Equal Power Distribution 110 Average Best Throughput post 4000 m

320 100 1000

Average Best Throughput till 4000 m

320 800 1000

Average no. of users

320 1000 0

200

Modified Water Filling

400 Water Filling

600

800

1000

1200

Equal Power

Fig. 5 A comparative analysis of the modified water-filling algorithm, the traditional water-filling algorithm, and the equal power distribution algorithm reveal distinct approaches in power allocation strategies [22]

Drones as an Alternate Communication System During Calamities

117

fosters efficient and optimal communication in disaster-affected zones by synergizing the PSO algorithm for drone placement and the water-filling algorithm for power allocation. These zones necessitate dependable communication for practical rescue and relief endeavors. Taken as a whole, the findings, coupled with judicious employment of optimization and power allocation methodologies, contribute to informed decision-making and the identification of prime strategies for deploying drones as relay media in disaster-prone or impacted regions.

6 Technical Parameters for Simulation The simulation parameters established for drawing conclusions are detailed in Table 1. The algorithmic parameters established for drawing conclusions are detailed in Table 2. Table 1 Simulation parameters [22] Measurements of area R

(X 1 , X 2 , Y 1 , Y 2 )

The highest quantity of impacted users resulting from M a disaster

(0,1000,0,1000) 100–1000

The maximum transmission capability of the drone

P (drone)

30 mW

Maximum transmission capacity for the affected users

P (user)

46 mW

Bandwidth of drone

BW (drone)

50 MHz

Bandwidth of user

BW (user)

75–100 MHz

Position of the distant communication base station

Loc (remote_base_stn) (7000, 500)

Table 2 Algorithmic parameters [22]

Frequency of the carrier signal Fc

2 GHz

PSO size

PSO (population) 50

Quantity of iterations

Iteration (PSO)

50

Arbitrary parameters

(a, b)

(9.2, 0.8)

Magnitude of noise power

Noise (power)

−120 mW

Threshold signal-to-noise

SNth

35 dB

118

D. S. Vohra et al.

7 Water-Filling Algorithm Utilizing a water-filling algorithm has demonstrated its efficacy in assessing optimal power distribution among drones and ground users. The arrangement of users within the impacted region “R” is depicted in Fig. 5. In this scenario, the aggregate transmission bandwidth amounts to “X” megahertz (MHz), with each user being allocated a fraction of “X/1000” bandwidth to ensure satisfactory transmission and reception capabilities. When considering a total bandwidth of 50 MHz and a user being assigned 50 Kbps, we assume this allocation allows for effective ground-user communication [14]. For comparison with alternative power allocation methodologies, the equal power allocation algorithms were also examined alongside the water-filling algorithm [1]. Findings indicated that the water-filling algorithm consistently surpasses the ergodic capacity exhibited by comparable communication channels [8]. The core concept encompassed identifying a threshold value dictating the point at which the drone should initiate transmission. The strategy involves channeling additional power to favorable communication channels between the drone and the ground user while still according to due consideration to unfavorable channels [12]. In this approach, attaining a balance between the judicious allocation of optimal power for favorable channels and the acknowledgment of unfavorable ones is paramount. In analyzing the water-filling algorithm, efforts were made to allocate power below the threshold value. However, transmitting negative power lacked practical significance and was thus excluded from consideration. The water-filling algorithm, in contrast to other power allocation techniques, secures heightened coverage for users, rendering it a preferred choice [15]. All in all, the adaptive nature of the water-filling algorithm empowers it to optimize power distribution, thereby fostering efficient communication and comprehensive coverage for ground users operating within the drone relay system (Fig. 6). However, as per the users consuming power till 4000 m, the Modified WaterFilling Algorithm performed very well. But here in this work, the concentration is on a scenario where there is a high probability that the remote communication base towers will be far off, more than 4000 m. The comparison between the modified water-filling algorithms and equal power distribution is summarized post-simulation as per the pictorial diagram in Fig. 7.

Fig. 6 Arrangement of users within the impacted subsection [22]

Drones as an Alternate Communication System During Calamities

119

Fig. 7 Power consumption versus users in a simulated environment till 4000 m. Source Prepared by authors

8 Results and Discussion The study’s outcomes unmistakably reveal the water-filling algorithm’s superior performance when distances exceed the 4000 m threshold. However, within the domain of 4000 m and lower distances, the advanced and equal power distribution algorithms exhibited better results than the water-filling algorithm. While these findings may yield theoretical insights, it is paramount to account for pragmatic considerations. As mentioned, disaster-affected regions typically span a breadth of 5000– 10,000 m in disaster-prone areas. In such contexts, the relevance of utilizing the waterfilling and equal power distribution algorithms becomes more pronounced. Given the expansive nature of disaster-hit territories, the practical implications derived from the research point toward the water-filling algorithm retaining greater relevance for situating drones as relay conduits within these areas. This research approach avoids drawing incorrect theoretical inferences by aligning with the practical dimensions of disaster-stricken zones. The spotlight is placed on the necessity of making astute decisions grounded in real-world circumstances to ensure the effective implementation of drone relay systems, serving communication and disaster response objectives. Acknowledging the practical realities surrounding disaster-hit areas is pivotal, as it underscores the significance of tailoring the selection of the apt power allocation algorithm according to the specific application and geographical context. This tactful approach guarantees the triumphant deployment of drone relay systems, adeptly addressing communication requisites within disaster-vulnerable regions.

120

D. S. Vohra et al.

9 Conclusion Leveraging drones as airborne relay mediums offers a promising solution for establishing connectivity in areas lacking pre-existing communication infrastructure or facing disruptions like disasters [7]. However, before embracing this concept, an intricate examination of path loss models becomes imperative, as these models constitute the foundational underpinning of any communication framework. The cellular-todrone path loss model was meticulously investigated to establish a backhaul link connecting the terrestrial base station with the drone. Moreover, to establish the downlink connection between the drone and ground users, an exhaustive exploration into the path loss model between the drone and the ground users was undertaken [6]. After the comprehensive analysis of both path loss models, a judicious equilibrium was achieved to discern the optimal positioning of drones. This intricate optimization dilemma, incorporating various constraints, was methodically formulated as a convex mathematical challenge. The primary goal encompassed enhancing the throughput for ground users by strategically locating drones as relays. In this context, the waterfilling power allocation algorithm proved apt for orchestrating effective drone placement, amplifying the system’s overall performance. Building upon this approach, a symmetrical exploration of the reverse scenario involving the uplink communication link—where ground users engage with drones—can also be scrutinized meticulously [24]. By factoring in the communication link between ground users and drones, a holistic evaluation of the complete drone relay system can be executed, effectively guaranteeing seamless bidirectional communication connecting ground users and the base station, all orchestrated through strategically positioned drones. This research underscores the pivotal role played by path loss models, optimization methodologies, and power allocation algorithms in crafting efficient drone relay systems. These systems bear the potential to extend communication coverage, amplify connectivity, and yield substantial benefits across diverse scenarios.

References 1. Al-Hourani A, Gomez K (2017) Modeling cellular-to-UAV path-loss for suburban environments. IEEE Wirel Commun Lett 7(1):82–85 2. Al-Hourani A, Kandeepan S, Jamalipour A (2014) Modeling air-to-ground path loss for low altitude platforms in urban environments. In: IEEE global communications conference, Austin, TX, USA, pp 2898–2904 3. Al-Hourani A, Kandeepan S, Lardner S (2014) Optimal lap altitude for maximum coverage. IEEE Wirel Commun Lett 3(6):569–572 4. Alnagar SI, Salhab AM, Zummo SA (2020) Unmanned aerial vehicle relay system: performance evaluation and 3D location optimization. IEEE Access 8:67635–67645 5. Chen Y, Feng W, Zheng G (2017) Optimum placement of UAV as relays. IEEE Commun Lett 22(2):248–251 6. Demir U, Toker C, Ekici O (2020) Energy-efficient deployment of UAV in v2x network considering latency and backhaul issues. In: IEEE international Black Sea conference on communications and networking (Black Sea Com), Odesa, Ukraine, pp 1–6

Drones as an Alternate Communication System During Calamities

121

7. Garg PK (2021) Introduction to unmanned aerial vehicles. New Age International Pvt Ltd, Delhi 8. He Y, Zhai D, Wang D, Tang X, Zhang R (2020) A relay selection protocol for UAV-assisted VANETs. Appl Sci 10(23):876 9. Hildmann H, Kovacs E (2019) Using unmanned aerial vehicles (UAVs) as mobile sensing platforms (MSPs) for disaster response, civil security, and public safety. Drones 3(3):59 10. Kalantari E, Yanikomeroglu H, Yongacoglu A (2016) On the number and 3D placement of drone base stations in wireless cellular networks. In: 2016 IEEE 84th vehicular technology conference (VTC-Fall), Montreal, QC, Canada, pp 1–6 11. Khuwaja AA, Chen Y, Zhao N, Alouini MS, Dobbins P (2018) A survey of channel modeling for UAV communications. IEEE Commun Surv Tutor 20(4):2804–2821 12. Kumar S, Suman S, De S (2018) Backhaul and delay-aware placement of UAV-enabled base station. In: IEEE INFOCOM 2018-IEEE conference on computer communications workshops (INFOCOM WKSHPS), Honolulu, HI, USA, pp 634–639 13. Mozaffari M, Saad W, Bennis M, Debbah M (2015) Drone small cells in the clouds: design, deployment and performance analysis. In: 2015 IEEE global communications conference (GLOBECOM), San Diego, CA, USA, pp 1–6 14. Palomar DP, Fonollosa JR (2005) Practical algorithms for a family of water filling solutions. IEEE Trans Signal Process 53(2):686–695 15. Perera TDP, Jayakody NK, Garg S, Kumar N, Cheng L (2020) Wireless-powered UAV assisted communication system in Nakagami-m fading channels. In: IEEE 17th annual consumer communications & networking conference (CCNC), Las Vegas, NV, USA, pp 1–6 16. Sawalmeh AH, Othman NS, Shakhatreh H, Khreishah A (2017) Providing wireless coverage in massively crowded events using UAVs. In: 2017 IEEE 13th Malaysia international conference on communications (MICC), Johor Bahru, Malaysia, pp 158–163 17. Sawalmeh AH, Othman NS, Shakhatreh H, Khreishah A (2019) Wireless coverage for mobile users in dynamic environments using UAV. IEEE Access 7:126376–126390 18. Shakhatreh H, Sawalmeh AH, Al-Fuqaha A (2019) Unmanned aerial vehicles (UAVs): a survey on civil applications and key research challenges. IEEE Access 7:48572–48634 19. Shakoor S, Kaleem Z, Baig MI, Chughtai O, Duong TQ, Nguyen LD (2019) Role of UAVs in public safety communications: energy efficiency perspective. IEEE Access 7:140665–140679 20. Ullah H, Nair NG, Moore A, Nugent C, Muschamp P, Cuevas M (2019) 5G communication: an overview of vehicle-to-everything, drones, and healthcare use-cases. IEEE Access 7:37251– 37268 21. Ullah Z, Al-Turjman F, Mostarda L (2020) Cognition in UAV aided 5g and beyond communications: a survey. IEEE Trans Cogn Commun Netw 6(3):872–891 22. Vohra DS, Garg PK, Ghosh SK (2023) Communication via quad/hexacopters during disasters. Int J Intell Unmanned Syst 23. Yan C, Fu L, Zhang J, Wang J (2019) A comprehensive survey on UAV communication channel modeling. IEEE Access 7:107769–107792 24. Yin S, Qu Z, Li L (2018) Uplink resource allocation in cellular networks with energyconstrained UAV relay. In: 2018 IEEE 87th vehicular technology conference (VTC Spring), Porto, Portugal, pp 1–5 25. Zeng Y, Wu Q, Zhang R (2019) Accessing from the sky: a tutorial on UAV communications for 5G and beyond. Proc IEEE 107(12):2327–2375 26. Zhan P, Yu K, Swindlehurst AL (2011) Wireless relay communications with unmanned aerial vehicles: performance and optimization. IEEE Trans Aerosp Electron Syst 47(3):2068–2085 27. Zhang L, Zhao H, Hou S (2019) A survey on 5g millimeter-wave communications for UAVassisted wireless networks. IEEE Access 7:117460–117504

Drought Analysis of an Area Using Google Earth Engine Jyothsna Devi Adapa and Keesara Venkatareddy

Abstract Drought is a long period of low rainfall which affects the growth of plants and living organisms. Drought occurs when an area or region suffers below-average rainfall or a water deficit for a prolonged time period. It is one of the most complex environmental disasters across the world. Drought also creates the same impact like floods or cyclones. Drought is categorized as meteorological, hydrological, agricultural and socio-economic. The area which experiences the precipitation deficit subjects to meteorological drought. Reduction in stream flow, ground water, and reservoir and lake levels indicates Hydrological drought. Soil water depletion causes Agricultural drought. The effect caused by the meteorological, hydrological and agricultural drought on people and economic activities is termed as socio-economic drought. This study is focusing on meteorological drought. Standard Precipitation Index (SPI) and Drought Severity Index (DSI) are used to calculate meteorological drought. Keywords Drought · SPI · Precipitation · DSI

1 Introduction 1.1 Drought When an area or region experiences abnormally low rainfall or a shortage of water for a prolonged period of time, then the area is said to be under drought. The amount of rainfall of an area varies from year to year. The lack of rainfall reduces the ground water or soil moisture. Due to reduction of ground water or soil moisture, there is crop damage. It takes weeks or months to identify drought. We can’t say exactly how long the drought lasts. It may be for weeks, months or years.

J. D. Adapa (B) · K. Venkatareddy Department of Civil Engineering, National Institute of Technology, Warangal, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_10

123

124

J. D. Adapa and K. Venkatareddy

1.2 Drought in India India has been witnessing drought since the eighteenth century. About 28% of the geographical area of India is subject to drought. From 1870 to 2018, India witnessed 26 major droughts out of which the drought that occurred in 2016–2018 was severe [1]. The drought in India is due to the delay of monsoon or failure of monsoon. During drought, India witnessed several deaths due to hunger and even suicides too. In order to avoid these incidents in future, drought monitoring is necessary.

1.3 About Google Earth Engine Google Earth Engine (GEE) is a cloud-based geospatial analysis platform that enables users to visualize and analyze satellite images of our planet. GEE uses JavaScript or Python programming language. Deforestation detection, Identification of drought, classifying land cover and Flood monitoring are some of the applications of GEE. GEE helps us to analyze the future conditions based on the past available data. GEE contains parameters which are useful for the assessment of drought like precipitation, temperature, evapotranspiration, soil moisture and vegetation. With the help of these parameters, we can calculate the standard precipitation index (SPI) and Drought Severity Index (DSI) of a particular area.

2 Literature Review Pai et al. [2] conducted a study to determine district-wide drought climatology in India over south-west monsoon season (June–September) by calculating the Percent of Normal Precipitation (PNP) and SPI for the period 1901–2003. In this study, rainfall data of all 458 districts in India is used. This study concluded that SPI is a better drought index as compared to PNP for calculating drought as PNP depends on the aridity of the region. Jahangir Alam et al. [3] conducted a study to monitor meteorological and agricultural drought in Bangladesh, Barind region using SPI and Markov method. They collected the 12 stations’ rainfall data from 1971 to 2008. Using gamma function, the precipitation data was normalized to compute SPI. They concluded that meteorological drought existed in the area in 1982, 1994 and 2010. Among these, the drought which occurred on 2010 was devasting. In a span of 5 years (2006–2010), meteorological drought occurred in the Barind region four times. Shah et al. [4] conducted a study for determining drought in Surat, Gujarat by calculating the SPI index. Forty years of precipitation data was collected for this study i.e., from 1971 to 2010 and used gamma distribution method to find SPI values. They assessed that droughts of moderate to exceptional severity occurred in 1974, 1985,

Drought Analysis of an Area Using Google Earth Engine

125

1987, 1989, 1991, and 1995. 1976, 1983, 1988, 2003, 2004, 2005, and 2007 were all judged to be moderately to highly wet. Surat district was determined to be in normal condition in the remaining years. Sazib et al. [5] conducted an experimental study to find out drought characteristics and comparison between meteorological and agricultural drought indicators using GEE. This study was conducted in Ethiopia and South Africa. The satellitebased soil moisture datasets were used to calculate surface and Root Zone Soil Moisture (RZSM) and their anomalies. Standard Precipitation Index (SPI) i.e., SPI3, SPI6, SPI9 was calculated. They concluded that in comparison to meteorological drought indicators, RZSM anomalies have longer drought duration but lower drought intensity. Zhao et al. [6] conducted an experimental study to examine the drought status and the trend change of Yellow River Basin from 2003 to 2019 in GEE. In this study, Moderate Resolution Imaging Spectroradiometer (MODIS) is used as a data source. MODIS Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI) are used to calculate the Temperature Condition Index (TCI), Vegetation Condition Index (VCI), Vegetation Health Index (VHI), and TemperatureVegetation Drought Index (TVDI). In high vegetation coverage areas, NDVI is more sensitive to drought chances so, for such regions drought monitoring through VCI is more preferable. In deserts, drought monitoring through TCI is better. They concluded that the TCI and TVDI show almost the same drought characteristics and VCI and VHI exhibit same drought characteristics. Khan and Gilani [7] conducted a study to determine drought of North America, South America, Africa, Europe, Asia, and Australia from 2001 to 2019 using GEE. In this study, they used MODIS satellite terra sensor evapotranspiration (ET), potential evapotranspiration (PET) 8-Day Global 500 m and MODIS Terra NDVI 8-Day Global 500 m data. They concluded that Australia, Africa, and Asia are undergoing frequent and extreme drought events. Europe is the least affected when compared to other continents. Khan and Gilani [8] conducted a study for drought monitoring using GEE in North America, South America, Europe, Asia, Africa, and Australia from 2001 to 2019. In this study, NDVI, LST, Soil moisture and Precipitation and Land cover mask data were used. With these data Vegetation Condition Index (VCI), Temperature Condition Index (TCI), Soil Moisture Condition Index (SMCI), and Precipitation Condition Index (PCI) were computed. VCI is used for agricultural drought, PCI is used for meteorological drought, and SMCI is used for agricultural and hydrological drought. They concluded that every continent is indicating signs of severe to extreme drought.

126

J. D. Adapa and K. Venkatareddy

Table 1 Condition of an area with SPI values (NDMC)

SPI values

Condition of the area

≤−2

Extremely dry

−1.5 to −1.99

Severely dry

−1.0 to −1.49

Moderately dry

−0.99 to +0.99

Near normal

1.0–1.49

Moderately wet

1.5–1.99

Very wet

>2.0

Extremely wet

3 Procedure 3.1 Procedure for Calculation of SPI Firstly, the monthly average precipitation data for 2000–2021 is collected from GEE using CHIRPS dataset. Using Excel, Mean and Standard deviation of the precipitation data is computed. Using Eq. (1) the value of SPI is calculated. The condition of the area in every month of a year is determined with SPI values from Table 1 SPI =

X i − X avg X sd

(1)

where X i = Precipitation data for the particular month, X avg = Mean/Average Rainfall, X Sd = Standard Deviation of Rainfall.

3.2 Procedure for Calculation of DSI Firstly, the data of ET and PET is collected from 2000 to 2021 from GEE using MODIS Global Terrestrial Evapotranspiration 8-Day Global 1 km. Using Excel, ratio of ET and PET is calculated and is mentioned as transpiration ratio (tran). Z1 is calculated which is the normalized value of tran. The equation for Z1 is as follows: Z 1 = (trani − tranmean )/transd

(2)

where trani = transpiration ratio for a particular month, tranmean = average transpiration ratio and transd = standard deviation of transpiration ratio. NDVI data is obtained from MODIS Terra Vegetation Indices 16-Day Global 500 m in GEE. The normalized value of NDVI is Z2, the calculation of which is done in Excel. The equation of Z2 is as follows: Z 2 = (NDVIi − NDVImean )/NDVIsd

(3)

Drought Analysis of an Area Using Google Earth Engine

127

where NDVIi = NDVI for a particular month, NDVImean = average NDVI and NDVIsd = Standard deviation of NDVI. The value Z is the summation of Z1 and Z2 Z = Z1 + Z2

(4)

DSI is the normalized value of Z. The equation for the calculation of DSI is as follows: DSI =

Z i − Z mean Z sd

(5)

where Z i = Z for a particular month, Z mean = average value of Z and Z sd = standard deviation of Z.

4 Database Collection from GEE Jangaon, Karimnagar, Khammam, Mahbubnagar, and Nizamabad shape files were independently collected and uploaded to GEE. Climate Hazardous Group Infrared Precipitation with Station data (CHIRPS) was used to collect precipitation for the above mentioned districts in the form of pentad. Figures 1 and 2 show the research area’s boundaries and satellite imagery. For the above districts, total rainfall and mean rainfall were calculated and plotted from 2000 to 2020. The monthly average rainfall for Jangaon district is shown in Fig. 3. The collection of data from NDVI, ET and PET for Jangaon district are shown in Figs. 4, 5, and 6 respectively. Fig. 1 Study area boundaries

128

J. D. Adapa and K. Venkatareddy

Fig. 2 Satellite image of study area

Fig. 3 Mean monthly rainfall for Jangaon district

Fig. 4 NDVI data for Jangaon district

NDVI of Jangaon District

0.8

NDVI

0.4 0.2

TIME

April-19

January-21

July-17

January-14

October-15

July-10

April-12

October-08

April-05

January-07

July-03

October-01

0 January-00

NDVI

0.6

Drought Analysis of an Area Using Google Earth Engine Fig. 5 ET data for Jangaon district

129

ET of Jangaon District

150

ET ET

100 50

April-20

January-18

October-15

July-13

April-11

January-09

July-04

October-06

April-02

January-00

0

TIME

Fig. 6 PET data for Jangaon district

1500

PET of Jangaon District PET

PET

1000 500

April-19

January-21

July-17

October-15

January-14

April-12

July-10

October-08

January-07

April-05

July-03

January-00

October-01

0

TIME

5 Results and Discussions Based on SPI values, the condition of the area is classified into seven categories. Near normal condition indicates that the particular area is neither subjected to drought nor received excess rainfall and for this condition, SPI values lies between −0.99 and +0.99. SPI values greater than 0.99 indicate wet condition and the value greater than 1.99 indicates extremely wet condition of the area. SPI values less than −0.99 indicates dry condition of the area and a value less than −1.99 indicates extremely dry condition of the area. Based on DSI values, the condition of the area is divided into eleven categories. For near normal conditions, DSI values lie in the range of 0.49 to −0.49. DSI values greater than 0.49 indicate the wet condition and a value greater than 3.99 indicates extremely wet condition of the area. DSI values less than −0.49 indicates dry condition of the area and the value less than −3.99 indicates extremely dry condition of the area.

130

J. D. Adapa and K. Venkatareddy

SPI values for Jangaon SPI

April-20

January-18

October-15

July-13

April-11

January-09

October-06

-5

July-04

0 January-00

SPI Values

5

April-02

Fig. 7 SPI values for Jangaon district

Time Period

5.1 Jangaon District The condition of the area based on SPI values for Jangaon district from 2000 to 2021 for all months is tabulated in Table 1. From the results, it is concluded that the Jangaon district was in moderately dry condition in April 2002, 2016, May 2003, 2008, 2015, June 2011, 2014, July 2001, 2002, 2009, 2019, August 2013, 2015, 2016, September 2000, 2011 and in October 2015. The district has experienced severe drought in July 2015, August 2004, and September 2002. The SPI values for Jangaon district from 2000 to 2021 are shown in Fig. 7. The condition of an area based on DSI values for Jangaon district from 2000 to 2021 is shown in Table 2. From the results, it is observed that the Jangaon district has Mild drought in January 2001, 2003, 2012, in February 2016, 2017, in March 2016, 2018, 2019, in April 2009, 2020, 2021, in May 2011, 2012, 2013, in June 2001, 2002, 2012–2014, in July 2002, 2015, 2016, in August 2005, in September 2018, 2019, in October 2008, 2018, in November 2000, 2009, 2011. The district has experienced moderate drought in September 2002 and October 2009. The DSI values for Jangaon district are shown in Fig. 8 (Tables 3 and 4).

5.2 Karimnagar District The Karimnagar district was in moderately dry condition in May 2003, 2007, 2012, June 2011, 2014, July 2009, August 2005, 2013, 2015, 2016, September 2000, 2003, October 2006 and 2015 whereas it experienced severe drought in July 2002 and September, August 2004, July 2015. The SPI values for Karimnagar district are shown in Fig. 9. The Karimnagar district is in mild drought condition in January 2001, 2004, 2006, 2012, in February 2005, 2010, 2016, in March 2005, 2007, 2010, 2016, in April 2003, 2005, 2010, 2012, in May 2001, 2005, in June 2001, 2002, 2006, in July 2011, in August 2001, 202, 2004, 2016, in September 2000, 2002, 2008, 2009, in October

Drought Analysis of an Area Using Google Earth Engine Table 2 Condition of an area with PDSI values (NDMC)

Fig. 8 DSI values for Jangaon district

131

PDSI values

Condition of the area

≤−4.0

Extreme drought

−3.0 to −3.99

Severe drought

−2.0 to −2.99

Moderate drought

−1.0 to −1.99

Mild drought

−0.5 to −0.99

Incipient dry spell

0.49 to −0.49

Near normal

0.5–0.99

Incipient wet spell

1.0–1.99

Slightly wet

2.0–2.99

Moderately wet

3.0–3.99

Very wet

≥4.0

Extremely wet

10

DSIvaluesforJangaon

0

-5

February-00 September-01 April-03 November-04 June-06 January-08 August-09 March-11 October-12 May-14 December-15 July-17 February-19 September-20

DSI VALUES

5

TIME PERIOD

2008, 2018, in November 2002, 2004, 2007, 2008, 2015, 2018, in December 2002, 2004, 2015 and it has moderate drought in July 2000. The DSI values for Karimnagar district are shown in Fig. 10.

5.3 Mahbubnagar District Mahbubnagar district is in moderately dry condition in April 2000, May 2001, 2003 in April, May and September, October 2006, 2008 in April and May, July 2009, June 2011, 2012 in May and September, June 2014, July 2015, April 2016, April 2017, 2018 in July and September and in Severe drought in September 2002, August 2004 and September 2011. The SPI values for Mahbubnagar district are shown in Fig. 11. Mahbubnagar district had mild drought in January 2016, February 2001, 2005, March 2000, 2016, April 2001, 2003, 2005, 2019, May 2003, 2005, June 2002,

132

J. D. Adapa and K. Venkatareddy

Table 3 Condition of Jangaon district using SPI values Month

Condition of the area Extremely Very wet Moderately Near wet condition wet normal condition condition area condition

Moderately Severely Extremely dry dry dry condition condition condition

January

2005, 2020

February

2011

March

2008

April

2006

2013, 2015

2004, 2012 2000–2021 2002, 2016 (except 2002, 2004, 2006, 2012, 2013, 2015, 2016)

May

2006

2016

2000

2000–2021 2003, (except 2008, 2015 2000, 2003, 2006, 2008, 2015, 2016)

June

2007

2000

2015

2000–2021 2011, 2014 (expect 2000, 2007, 2011, 2014, 2015)

2019

2005, 2010

2000–2021 (except 2005, 2019, 2020) 2000–2021 (except 2005, 2010 and 2011) 2000–2021 (except 2008)

(continued)

Drought Analysis of an Area Using Google Earth Engine

133

Table 3 (continued) Month

Condition of the area Extremely Very wet Moderately Near wet condition wet normal condition condition area condition

July

August

2000

September

October

2020

Moderately Severely Extremely dry dry dry condition condition condition

2005, 2010

2000–2021 2001, 2015 (except 2002, 2001, 2009, 2019 2002, 2005, 2009, 2010, 2015, 2019)

2008, 2010

2007, 2020 2000–2021 2013, 2004 (except 2015, 2016 2000, 2004, 2007, 2008, 2010, 2013, 2015, 2016, 2020)

2016, 2020

2005

2000–2021 2000, 2011 2002 (except 2000, 2002, 2005, 2011, 2016, 2020)

2005, 2013

2019

2000–2021 2015 (except 2005, 2013, 2015, 2019, 2020) (continued)

134

J. D. Adapa and K. Venkatareddy

Table 3 (continued) Month

Condition of the area Extremely Very wet Moderately Near wet condition wet normal condition condition area condition

November 2010

2012

December 2010, 2018

Moderately Severely Extremely dry dry dry condition condition condition

2006, 2000–2021 2008, 2021 (except 2006, 2008, 2010, 2012, 2021) 2000–2021 (except 2010, 2018)

2003, 2011, July 2000, 2009, August 2001, 2002, 2006, 2016, September 2001, 2002, 2008, 2011, October 2004, 2006, 2016, November 2004, 2015, 2018 and in December 2000, 2004, 2015 and it has experienced moderate drought in May 2001. The DSI values for Mahbubnagar district are shown in Fig. 12.

5.4 Nizamabad District The Nizamabad district is in Moderately dry condition in September 2000, July 2001, September 2002, 2003 in May and September, 2004 in June and August, October 2006, May 2007, June 2008, July 2009, June 2011, May 2012, August 2013, 2017 in April and July, September 2018 and the district has experienced severe dry condition in July 2002, June 2014, July 2015 and August 2016. The SPI values for Nizamabad district are shown in Fig. 13. The Nizamabad district has experienced mild drought in January 2003, 2016, February 2003, 2004, 2016, March 2016, April 2003, 2010, 2016, 2019, May 2001, 2019, June 2001, 2002, 2008, July 2002, 2011, August 2002, 2004, 2010, 2013, September 2000, 2003, 2015, October 2000, 2014, 2015, 2018, November 2000 and in December 2000, 2002, 2015. The district was in moderate drought conditions in July 2000 and August 2016. The DSI values for Nizamabad district are shown in Fig. 14.

2013, 2015

2006, 2016

April

May

June

2008

Moderately wet

March

2020

Very wet

2010, 2011

Extremely wet

Condition of the area

February

January

Month

2000, 2007, 2015

2002, 2004

2012

2011

2008, 2020, 2021

2005, 2019

Slightly wet

2016, 2017, 2018

2003, 2014, 2015

2002, 2006

2000, 2001, 2014, 2021

2007, 2021

Incipient wet spell

Table 4 Condition of an area based on DSI values for Jangaon district

2004, 2005, 2006, 2003, 2009, 2008, 2019, 2020, 2010, 2011 2021

2000, 2001, 2005, 2007, 2018, 2008, 2009, 2010, 2019, 2020, 2017 2021

2000, 2001, 2003, 2008, 2010, 2004, 2005, 2007, 2017, 2018, 2011, 2014, 2016 2019

2002, 2006, 2010, 2003, 2004, 2012, 2013, 2020 2005, 2007, 2009, 2015, 2017

2001, 2002, 2005, 2003, 2004, 2006, 2009, 2012, 2007, 2015, 2013, 2014, 2019 2018

2001, 2002, 2012, 2013, 2014

2011, 2012, 2013

2009, 2020, 2021

2016, 2018, 2019

2016, 2017

2001, 2003, 2012

Incipient dry Mild spell drought

2002, 2004, 2006, 2000, 2015, 2008, 2009, 2010, 2016 2011, 2014, 2017, 2018, 2020

Near normal

Moderate drought

Severe drought

(continued)

Extreme drought

Drought Analysis of an Area Using Google Earth Engine 135

2008, 2010, 2021

2007, 2020

2001, 2005, 2010, 2013, 2016, 2020 2013, 2020

August

September

October

2000, 2001, 2012, 2019

2009, 2021

2008, 2018, 2019, 2020, 2021

2012

Incipient wet spell

2010

July

Very wet

Slightly wet

Extremely wet

Condition of the area

Moderately wet

Month

Table 4 (continued)

2002, 2003, 2004, 2006, 2007, 2005, 2010, 2011, 2021 2014, 2015, 2016, 2017

2000, 2003, 2004, 2006, 2007, 2008, 2012, 2014, 2011, 2015 2017

2000, 2001, 2002, 2003, 2004, 2009, 2011, 2012, 2006, 2013, 2014, 2019 2015, 2016, 2017, 2018

2008, 2018

2018, 2019

2005

2002, 2015, 2016

Incipient dry Mild spell drought

2000, 2001, 2004, 2003, 2011, 2005, 2006, 2007, 2013, 2014 2009, 2017

Near normal

2009

2002

Moderate drought

Severe drought

(continued)

Extreme drought

136 J. D. Adapa and K. Venkatareddy

December

November

Month

Extremely wet

Very wet

Condition of the area

Table 4 (continued)

2021

Moderately wet

2005, 2018, 2019

2003, 2012

Slightly wet

2003, 2004, 2006, 2017

2008, 2015, 2017

Incipient wet spell

2007, 2009, 2016, 2002, 2008, 2020, 2021 2010, 2011, 2012, 2014, 2015

2000, 2009, 2011

Incipient dry Mild spell drought

2004, 2006, 2010, 2001, 2002, 2013, 2014, 2016, 2005, 2007, 2019, 2020 2018

Near normal

Moderate drought

Severe drought

Extreme drought

Drought Analysis of an Area Using Google Earth Engine 137

138

J. D. Adapa and K. Venkatareddy

Fig. 9 SPI values for Karimnagar district

SPI Values for Karimnagar

6

SPI

2 January-21

July-17

April-19

October-15

April-12

January-14

January-07

July-10

July-03

October-08

-4

April-05

-2

October-01

0 January-00

SPI Values

4

Time Period

Fig. 10 DSI values of Karimnagar district

DSI of Karimnagar District

4

DSI

August-20

January-21

June-17

January-19

April-19

April-14

November-15

September-12

July-09

February-11

May-06

-4

December-07

October-04

August-01

-2

March-03

0 January-00

DSI Values

2

Time Period

Fig. 11 SPI values for Mahbubnagar district

SPI for Mahbubnagar

5

July-17

October-15

April-12

January-14

July-10

October-08

April-05

July-03

January-07

-5

October-01

0 January-00

SPI Values

SPI

Time Period

5.5 Khammam District Khammam district is in moderately dry condition in September 2000, 2001 in May and July, April 2002, May 2003, May 2007, 2008 in April, May, September, 2009 in July and October, 2011 in June and September, August 2013, April 2014, 2015 in

DSI VALUES 3

-1

-2

-3

TIME PERIOD

June-17 January-19 August-20

4

November-15

Fig. 14 DSI values of Nizamabad district

July-09 February-11 September-12 April-14

-3 July-03 January-07

4

January-21

April-19

July-17

October-15

January-14

April-12

July-10

October-08

Fig. 13 SPI values of Nizamabad district

April-05

-4

January-20

January-18

January-16

January-14

January-12

January-10

January-08

January-06

January-04

January-02

January-00

4

October-04 May-06 December-07

-2

January-00

-1

-2

October-01

DSI values

Fig. 12 DSI values of Mahbubnagar district

January-00 August-01 March-03

SPI values

Drought Analysis of an Area Using Google Earth Engine 139

DSI of Mahbubnagar DSI

2

0

Time Period

SPI of Nizamabad SPI

3

2

1

0

Time period

DSI of Nizamabad DSI

2

1

0

140 Fig. 15 SPI values of Khammam district

J. D. Adapa and K. Venkatareddy

SPI of Khammam

5

SPI values January-00

spi

April-19

January-21

July-17

October-15

April-12

January-14

July-10

April-05

January-07

July-03

October-08

-5

October-01

0

Time Period

Fig. 16 DSI values of Khammam district

DSI of Khammam

5

DSI values

DSI

January-21

April-19

July-17

January-14

October-15

April-12

July-10

October-08

January-07

April-05

July-03

October-01

-5

January-00

0

Time Period

May, August and October, 2016 in April, August and October, April 2017 and Severe dry condition in September 2002, August 2004, June 2014 and in extreme drought in July 2002 and 2015. SPI values for Khammam district is shown in Fig. 15. Khammam district is in mild drought conditions in July 2000, 2001 in February and August, 2002 in April, August, September, November and December, 2003 in February, March, April, May, 2004 in August and October, 2005 in May and June, Jan 2007, June 2008, 2009 in September and October, 2011 in July and November, 2012 in February and March, 2016 in January, March, April and August, Oct 2018, September 2019 and has experienced moderate drought in May 2001. DSI values for Khammam district are shown in Fig. 16.

6 Summary and Conclusions GEE has user friendly environment where accessing the data is easy and fast. The aim of this research is to determine the area’s meteorological drought. From 2000 to 2021, the SPI and DSI figures for Jangaon, Karimnagar, Mahbubnagar, Nizamabad, and Khammam were calculated. Based on SPI and DSI values the condition of the

Drought Analysis of an Area Using Google Earth Engine

141

district is determined from Tables 1 and 2. The condition of the area based on SPI and DSI values is almost the same, but the SPI values which are close to the range are changing in DSI. Based on observations from our study, the district which is getting more repeated droughts needs immediate water resource management. The available water resources can be augmented by reservation of the existing water bodies and transferring water from the nearest reservoirs which are from Godavari and Krishna basins.

References 1. https://india.mongabay.com/2021/05/southern-indias-2016-2018-drought-was-the-worst-in150-years/ 2. Pai DS, Sridhar L, Guhathakurta P, Hatwar HR (2011) District-wide drought climatology of the southwest monsoon season over India based on standardized precipitation index (SPI). Nat Hazards. https://doi.org/10.1007/s11069-011-9867-8 3. Jahangir Alam ATM, Sayedur Rahman M, Saadat AHM (2013) Monitoring meteorological and agricultural drought dynamics in Barind region Bangladesh using standard precipitation index and Markov chain model. Int J Geomat Geosci 3(3). ISSN 0976-4380 4. Shah R, Bharadiya N, Manekar V (2015) Drought index computation using standardized precipitation index (SPI) method for Surat District, Gujarat. Aquat Procedia 4:1243–1249. https://doi. org/10.1016/j.aqpro.2015.02.162 5. Sazib N, Mladenova I, Bolten J (2018) Leveraging the Google Earth Engine for drought assessment using global soil moisture data. Remote Sens 10:1265. https://doi.org/10.3390/rs1008 1265 6. Zhao X, Xia H, Pan L, Song H, Niu W, Wang R, Li R, Bian X, Guo Y, Qin Y (2021) Drought monitoring over Yellow River Basin from 2003–2019 using reconstructed MODIS land surface temperature in Google Earth Engine. Remote Sens 13:3748. https://doi.org/10.3390/rs13183748 7. Khan R, Gilani H (2021) Global drought monitoring with drought severity index (DSI) using Google Earth Engine. Theor Appl Climatol 146:411–427. https://doi.org/10.1007/s00704-02103715-9 8. Khan R, Gilani H (2021) Global drought monitoring with big geospatial datasets using Google Earth Engine. Environ Sci Pollut Res 28:17244–17264. https://doi.org/10.1007/s11356-020-120 23-0 9. https://drought.unl.edu/ranchplan/DroughtBasics/WeatherandDrought/MeasuringDrought. aspx

Effects of Urbanization on Land Use Land Cover of Warangal Region Using RS and GIS Ch. Sree Laxmi Pavani, Keesara Venkatareddy, and S. Joshmitha

Abstract Rapid expansion of urbanization has increased the population and economic growth of towns and cities in various parts of the country. This increase in urbanization has affected the natural resources such as vegetation and water bodies. In the present study, the effects of urbanization on LULC (Land Use Land Cover) changes of Warangal urban and rural areas are studied for the years 2014, 2017 and 2020 for a span of 6 years using the RS and GIS techniques. LANDSAT imagery of 3 years is collected and enhanced and LULC data is extracted for four classes i.e., urbanization (built-up land), agriculture cum forest (vegetation), bear soil (barren land) and water bodies. RS and GIS tools are used to compare and analyze the effect of urbanization. LULC feature extraction is made for urban and rural areas of Warangal region using supervised classification to study the growth of urban areas and its effect on LULC changes. The results indicate a drastic increase in urbanization and barren land, severe decrease in vegetation and a very slight increase in waterbodies. Increase in urbanization and decrease in agriculture cum forest indicates vegetation reduction and built-up area increment which directly affects the climate and environment. Warangal region is being affected due to the rising population and the improper usage of land. The quantitative results obtained explain the effect of urbanization on LULC changes of Warangal rural and urban areas, which helps in the best management and planning of land use for Warangal region. Keywords Classification · Land use land cover · RS and GIS · Urbanization · Warangal

Ch. Sree Laxmi Pavani (B) · K. Venkatareddy Civil Engineering Department, National Institute of Technology, Warangal, Telangana, India e-mail: [email protected] S. Joshmitha Civil Engineering Department, Kakatiya Institute of Technology and Sciences, Warangal, Telangana, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_11

143

144

Ch. Sree Laxmi Pavani et al.

1 Introduction 1.1 Effects of Urbanization on LULC Land Use Land Cover (LULC) changes are affected in the cities and towns due to Rapid Urbanization [1]. This occurs due to the relocation of people from rural areas to urban areas. Monitoring Land use Land cover changes regularly will help the urban planners and decision makers to overcome the urban sprawl problem and will help in sustainable development [2]. Since 1950, water usage has increased more than tripled globally and one out of every six persons does not have regular access to safe drinking water [3]. For any type of sustainable development program LULC changes on the earth’s surface are enormously important. Due to various human activities on earth’s surface undesirable changes in LULC have taken place which led to land degradation [4]. Expansion of urbanization is a historic and speedy conversion of human life on a universal scale, whereby rural tradition is being quickly replaced by urban civilization. Many people from rural areas relocate to the city in search of financial success and social flexibility. Though, the picture of urbanization growth is not as great as it appears to be. As a result of rapid industrial development, modernistic cities have evolved in an unplanned and disorganized manner. Cities in emerging nations become overcrowded and overpopulated for a variety of reasons, including migration and long-term increase in population [5]. Understanding the characteristics of landscape is made imaginable by remote sensing techniques such as digital change detection using multitemporal Landsat-imagery. Remote sensing and GIS methods are essential technologies for the temporal quantification and investigation of spatial trends or else it is impossible when compared with traditional mapping methods. With the aid of RS and GIS technologies, change detection enables users with faster, affordable and more accurate results [6]. Visual interpretation provides a broad overview of the land cover changes and its classes across the specified time period. The effectiveness of change detection from images may vary depending on the type of change involved and the efficiency of the image pre-processing as well as the classification operations [7]. Using remote sensing, characterizing natural and urban ecosystems is important for change detection techniques. Accurate radiometric and geometric rectification is crucial for the analysis of changes in land use and land cover using multitemporal remote sensing data [8]. The worrying aspects are that land degradation is getting worse due to deforestation, farming on slopes and land fragmentation. GIS software served as a platform for data analysis and product management [9]. In the present study effect of urbanization on LULC of Warangal region of Telangana, India using RS and GIS techniques is carried out. Warangal rural and urban areas were considered to analyze the urbanization effects on land use land cover. The urban growth and its impacts are detected using RS and GIS tools and techniques. The study reveals that the major land covered in the study area is vegetation which is slowly being occupied by built-up and barren land. Effects of urbanization bring a lot of change in vegetation and water bodies which can lead to the occurrence of floods and fluctuations in ground water levels.

Effects of Urbanization on Land Use Land Cover of Warangal Region …

145

1.2 Study Area Study area map of Warangal region of Telangana is shown in Fig. 1. Warangal urban and rural areas are situated in Telangana, India with its geographical extent as 18° 16' 45.278'' to 17° 37' 4.0296'' N and 79° 13' 41.412'' to 80° 2' 28.478'' E. Warangal district was reformed in the year 2016 by carving out Warangal rural district from erstwhile Warangal district. Warangal Urban area was renamed in the year 2021 as Hanamkonda. Warangal Urban covers an area of 1294.4 km2 with 14 mandals and Warangal Rural covers an area of 2139 km2 with 13 mandals. Warangal region feels a dry and hot, seasonally wet, tropical climate. The average annual temperature of Warangal region is 27.2 °C. The monsoon season is from June to September and the rainfall recorded on an average is about 1020 mm.

2 Materials and Methods To obtain the Vector files of Warangal region, India and Telangana boundary shape files are collected to extract Warangal urban and rural areas using GIS techniques. The Landsat images are collected from the USGS Earth Explorer from band1 to band7 and are combined using the composite band tool in ArcGIS. To enhance the image, supervised classification is performed using Remote Sensing techniques. Supervised classification is performed by selecting training areas for each land cover class to be created. Signature file is generated with the collected training sample data. Now, using GIS signature file is taken as input to classify the study area. In Classify tool, maximum likelihood supervised classification method is applied to create LULC maps. The LULC maps are classified as urbanization, waterbodies, agriculture cum

Fig. 1 Study area map showing Warangal urban and rural regions located in Telangana, India

146

Ch. Sree Laxmi Pavani et al.

Fig. 2 Methodology of LULC change analysis

forest and bear soil for Warangal urban and rural areas. The changes are detected for 3 recent years 2014, 2017 and 2020 to compare the effect of urbanization on LULC. Methodology followed for the study area is shown in Fig. 2. Urbanization of Warangal urban and rural areas for the years 2014, 2017 and 2020 is shown in Figs. 3 and 4. In 2014, 2017 and 2020 urbanization occupied an area of 21.61%, 28.73% and 32.65% of the total area. This shows how rapidly the urbanization has increased by 11.04% in Warangal urban area from 2014 to 2020 as shown in Fig. 3. In rural areas of Warangal region urbanization has covered an area of 10.66%, 14.31% and 18.76% in the years 2014, 2017 and 2020. These values indicate that urbanization is growing by 8.1% for rural areas of Warangal region also.

Fig. 3 Urban area in 2014, 2017 and 2020

Effects of Urbanization on Land Use Land Cover of Warangal Region …

147

Fig. 4 Rural area in 2014, 2017 and 2020

3 Results LULC maps of Warangal Urban and Rural areas for the years 2014, 2017 and 2020 are shown in Fig. 5. The results indicate that Urbanization has increased drastically, waterbodies increased in very less quantity and agriculture and forest areas decreased severely from 2014 to 2020.

3.1 LULC Details of Warangal Urban Region The total area of Warangal urban is 1294.4 km2 , out of which in the year 2014, urbanization covered 279.80 km2 , agriculture cum forest is 957.5 km2 , water bodies covered 44.27 km2 and bear soil is covered by 12.87 km2 . In the year 2017, urbanization increased from 279.80 to 371.97 km2 which indicates an increase of 7.12% from 2014 to 2017. Agriculture and forest have decreased from 957.5 to 814.45 km2 with a percentage decrease of 11.05%. Water bodies have a very minimal increase of 0.46% from 44.27 to 50.2 km2 and bear soil has increased from 12.87 to 57.82 km2 with a percentage increase of 3.44% from 2014 to 2017. In the year 2020, urbanization has increased to 422.74 km2 with a percentage increase of 3.92%. Agriculture cum forest has decreased to 726.41 km2 with a percentage decrease of 6.8%. Water bodies have a slight increase of 0.79% from 50.2 to 60.4 km2 . Bear soil has increased from 57.82 to 84.87 km2 with a percentage increase of 2.09%. LULC details of Warangal Urban area are classified along with its percentages are arranged in Table 1. Percentage increase and decrease of each LULC feature of Warangal Urban area are shown in Table 6.

148

Ch. Sree Laxmi Pavani et al.

Fig. 5 LULC map of Warangal urban and rural areas for the years 2014, 2017 and 2020

Table 1 Warangal urban LULC classified area and percentage of the 4 categories i.e., urbanization, agriculture cum forest and bear soil for 2014, 2017 and 2020 Bear soil (km2 )

Total area (km2 )

Year

Urbanization (km2 )

Agriculture cum forest (km2 )

Water bodies (km2 )

2014

279.80 (21.61%)

957.5 (73.96%)

44.27 (3.41%) 12.87 (1.02%)

1294.4

2017

371.97 (28.73%)

814.45 (62.91%)

50.2 (3.87%)

57.82 (4.46%)

1294.4

2020

422.74 (32.65%)

726.41 (56.11%)

60.4 (4.66%)

84.87 (6.55%)

1294.4

3.2 LULC Details of Warangal Rural Region The total area of Warangal Rural is 2139 km2 , from which in the year 2014, urbanization covered an area of 228.14 km2 , agriculture cum forest 1823.79 km2 , water bodies covered 69.34 km2 and bear soil covered 16.79 km2 . In the year 2017, urbanization

Effects of Urbanization on Land Use Land Cover of Warangal Region …

149

increased from 228.14 to 306.22 km2 which indicates a percentage increase of 3.65% from 2014 to 2017. Agriculture cum forest decreased from 1823.79 to 1594.47 km2 with a percentage decrease of 10.74%. Water bodies have a negligible decrease of 0.17% from 69.34 to 65.68 km2 and bear soil has increased from 16.79 to 172.13 km2 with a percentage increase of 7.26% from 2014 to 2017. In the year 2020, from 2017 to 2020 urbanization has enlarged to 401.39 km2 with a percentage increase of 4.45%. Agriculture cum forest has diminished to 1497.39 from 1594.47 km2 with a percentage decrease of 4.56%. Water bodies have a nominal increase of 1.03% i.e., from 65.68 to 87.91 km2 . Bear soil has decreased from 172.13 to 152.10 km2 with a percentage decrease of 0.93%. LULC details of Warangal Rural classified area along with its percentages are tabulated in Table 2. Percentage increase and decrease of each LULC feature of Warangal Rural area is shown in Table 7. Bar graph of Warangal urban and rural regions is shown in Fig. 6 for the years 2014, 2017 and 2020. From the graph, it is observed clearly that agriculture cum forest is decreasing and in urban areas, bear soil is increasing. On the other side, there is a very slight increase in water bodies, which may not be sufficient for the growing needs of the population. Table 2 Warangal rural classified area and percentage of the 4 categories i.e., urbanization, vegetation, bear soil for 2014, 2017 and 2020 Bear soil (km2 )

Total area (km2 )

Year

Urbanization (km2 )

Agriculture cum forest (km2 )

Water bodies (km2 )

2014

228.14 (10.66%)

1823.79 (85.29%)

69.34 (3.24%) 16.79 (0.78%)

2139

2017

306.22 (14.31%)

1594.47 (74.55%)

65.68 (3.07%) 172.13 (8.04%)

2139

2020

401.39 (18.76%)

1497.39 (69.99%)

87.91 (4.10%) 152.10 (7.11%)

2139

Fig. 6 Bar graph of Warangal urban and rural regions

150 Table 3 Population growth of Warangal region

Ch. Sree Laxmi Pavani et al.

Rainfall data of Warangal region Year

Annual rainfall (mm)

2014

740

2017

892

2020

1853

Source https://www.census2011.co.in and Telangana socioeconomic outlook 2015–2016

3.3 Population Growth of Warangal Table 3 indicates the population growth rate of Warangal region. As per the 2001 census the year 2001 population of Warangal region was 3,246,004 which increased to 3,512,576 in the year 2011 and as per Intensive Household Survey 2014, the population of Warangal region increased to 3,647,000. Sweeping increase in population has led to an increase in urbanization of Warangal region. This population growth not only affected urbanization but also decreased agriculture cum forest and increased bear soil of Warangal region.

3.4 Annual Rainfall of Warangal Annual rainfall data of Warangal region is displayed in Table 4. It is observed that from 2014 to 2020 the rainfall has increased drastically from 740 to 1853 mm. Because of which there is a slight increase in water bodies. Due to the increase in population, humidity and moisture content may increase, because of which evaporation may increase and thereby as part of the hydrological cycle rainfall may also increase. Tables 4 and 5 show annual and monthly rainfall data recorded for the years 2014, 2017 and 2020. Average Annual rainfall data of Warangal region is collected from the central ground water board (CGWB) website to analyze the recorded rainfall. Yearly annual rainfall data shows an increasing trend from 2014 to 2020 (Table 4). Annual rainfall recorded in the years 2014, 2017 and 2020 are 740 mm, 892 mm and 1853 mm respectively. From 2014 to 2017 there was a rainfall increase of 152 mm. Whereas from 2017 to 2020 there is a sudden increase of extremely heavy rainfall of 961 mm. Especially in the month of August 2020 (Table 5) rainfall has risen from 321.7 mm (July) to 810.2 mm (August). During this period, low lying areas of Warangal region were totally flooded due to very heavy rainfall. Because of which there was hefty loss of people and property. Therefore, proper measures are to be taken to control heavy rainfall and flooded areas. Therefore, heavy rainfall flooded water shall be guided in a proper sequence for storage of water. Table 5 shows the monthly rainfall information of Warangal region for the years 2014, 2017 and 2020 collected from CGWB. Hence, for a slight increase in water bodies, rainfall increase is one of the

Effects of Urbanization on Land Use Land Cover of Warangal Region … Table 4 Annual rainfall data of Warangal region for the years 2014, 2017 and 2020

151

Population growth of Warangal region Year

Population

2001

3,246,004

Growth rate (%)

2011

3,512,576

8.21

2014

3,647,000

3.82

Source http://cgwb.gov.in

Table 5 Monthly rainfall data of Warangal region for the years 2014, 2017 and 2020 Year Jan Feb

Mar Apr May

2014 0.5 0

42.1 5

2017 0

0

0

Jun

105.6 53.9

19.5 1

2020 0.8 14.3 12.1 16.6 28.4

226

Jul

Aug

Sep

Oct

Nov Dec Annual

172.1 201.7 118.3 29.5

11

0

740

197.4 171

0

0

892

156

121

206.4 321.7 810.2 240.4 191.5 10.5 0

1853

Source http://cgwb.gov.in

Table 6 Percentage change in Warangal urban region Year

Urbanization

Agriculture and forest

Waterbodies

Bear soil

2014–2017

Increased by 7.12%

Decreased by 11.05%

Increased by 0.46%

Increased by 3.44%

2017–2020

Increased by 3.92%

Decreased by 6.8%

Increased by 0.79%

Increased by 2.09%

Table 7 Percentage change in Warangal rural region Year

Urbanization

Agriculture and forest

Waterbodies

Bear soil

2014–2017

Increased by 3.65%

Decreased by 10.74%

Decreased by 0.17%

Increased by 7.26%

2017–2020

Increased by 4.45%

Decreased by 4.56%

Increased by 1.03%

Decreased by 0.93%

causes. Not only this, Telangana Government has implemented tank rejuvenation as part of mission Kakatiya, to improve the groundwater resources. This is also one of the reasons for the increase in water bodies. From the results of Tables 1 and 2, it is observed that, urban development has rapidly increased in the span of 6 years. Urbanization has led to a decrease in vegetation. Waterbodies increased in very less quantity which is less than 1% and is not sufficient for the growing needs of the Warangal population. Necessary measures are being taken to protect and preserve the agricultural area, forest and water bodies of Warangal region. The urbanization has increased very rapidly to 11.04% in Warangal

152

Ch. Sree Laxmi Pavani et al.

urban area and in rural areas increased to 8.1% in the span of six years (Tables 6 and 7). Due to increase in the urbanization agricultural and forest areas have decreased drastically by 17.85% in urban areas and in rural areas by 15.3%. In Warangal Urban 231.09 km2 of agricultural land cum forest area is decreased and in Warangal Rural 326.4 km2 of agricultural land cum forest area is decreased. Urbanization area of 142.94 km2 got increased in Warangal Urban and 173.25 km2 of Urbanization area got increased in Warangal Rural.

4 Conclusions The results indicate that the effect of urbanization on LULC of Warangal region is more which has to be planned efficiently. Urban development highly affected bear soil and agriculture cum forest both in rural and urban areas. Very little increase in water bodies is observed in the Warangal urban and rural areas, which will not be sufficient for the growing needs of the population. Increase in water bodies are observed due to increase in rainfall and implementation of Mission Kakatiya Scheme by the Telangana government. Proper planning in correlation with the growing population and LULC changes is to be made for both urban and rural areas of Warangal region. To know the effect of urbanization on LULC changes of Warangal using RS and GIS techniques is necessary. Therefore, measures shall be taken to overcome the problem of urbanization in the Warangal region. Effects of Urbanization on Land Use and Land Cover of Warangal Region Using RS and GIS techniques is important to understand the changes of Warangal land cover.

References 1. Mohan M, Pathan SK, Narendrareddy K, Kandya A, Pandey S (2011) Dynamics of urbanization and its impact on land-use/land-cover: a case study of megacity Delhi. J Environ Prot 2(09):1274 2. Moniruzzaman M, Roy A, Bhatt CM, Gupta A, An NTT, Hassan MR (2018) Impact analysis of urbanization on land use land cover change for Khulna City, Bangladesh using temporal landsat imagery. Int Arch Photogramm Remote Sens Spat Inf Sci 42(5):757–760 3. Jeykumar RKC, Chandran S (2019) Impact of urbanization on climate change and geographical analysis of physical land use land cover variation using RS-GIS. Global NEST J 21(2):141–152 4. Abd El-Kawy OR, Rød JK, Ismail HA, Suliman AS (2011) Land use and land cover change detection in the western Nile delta of Egypt using remote sensing data. Appl Geogr 31(2):483– 494 5. Jaysawal D, Saha S (2014) Urbanization in India: an impact assessment. Int J Appl Sociol 4(2):60–65 6. Rawat JS, Kumar M (2015) Monitoring land use/cover change using remote sensing and GIS techniques: a case study of Hawalbagh block, district Almora, Uttarakhand, India. Egypt J Remote Sens Space Sci 18(1):77–84 7. Shalaby A, Tateishi R (2007) Remote sensing and GIS for mapping and monitoring land cover and land-use changes in the Northwestern coastal zone of Egypt. Appl Geogr 27(1):28–41

Effects of Urbanization on Land Use Land Cover of Warangal Region …

153

8. Treitz P, Rogan J (2004) Remote sensing for mapping and monitoring land-cover and land-use change-an introduction. Prog Plan 61(4):269–279 9. Mhawish YM, Saba M (2016) Impact of population growth on land use changes in Wadi Ziqlab of Jordan between 1952 and 2008. Int J Appl Sociol 6(1):7–14

Effect of LULC Changes on Land Surface Temperature Rajashekar Kummari, Pavan Kumar Reddy Allu, Shashi Mesapam, Ayyappa Reddy Allu, Bhargavi Vinakallu, and Bhanu Prakash Ankam

Abstract The rapid urbanization of cities has led to significant human-induced alterations to the environment, characterized by the replacement of natural land cover with man-made structures such as concrete, bricks, asphalt, and metal. This transformation disrupts the natural processes of evapotranspiration, resulting in reduced cooling effects and increased heat storage within urban areas. This study involves the Land Surface Temperature (LST) variation over different classes of Land Use Land Cover (LULC) and the effect of change in LULC over the past two decades in and around Hyderabad City, Telangana, India. This is a Tier-1 Indian city experiencing notable LULC changes from the past two decades due to rapid Urbanization. This study examines how urbanization has affected the surface temperatures from the past twenty years, using Landsat 7 and Landsat 8. Spectral indices such a LULC, Radiance, Brightness temperatures and Land Surface Emissivity (LSE) have been generated and utilized for the generation of LST. A substantial change of vegetation increased by 28.92% from 2012 to 2022 period. And also, an urban area increasing by 108.8% but waterbodies have decreasing and increasing patterns in the period of 2012–2022. It is advised to implement efficient urban planning and design solutions to lessen the negative consequences of LST changes caused by LULC. Green space inclusion, vegetation restoration, cool pavement and roofing techniques, and sustainable urban design techniques are among those to be employed in order to reduce the effect of these temperature changes. Keywords Urbanization · Land Surface Temperature (LST) · Urban Heat Island (UHI) · Normalized Difference Vegetation Index (NDVI)

R. Kummari · P. K. R. Allu (B) · S. Mesapam · A. R. Allu · B. Vinakallu · B. P. Ankam Department of Civil Engineering, National Institute of Technology, Warangal, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_12

155

156

R. Kummari et al.

1 Introduction Rapid urbanization in the past decades has resulted in a great change in the Land Use Land Cover (LULC) and created a great impact on the environment and climate [1, 2]. One such important consequence of LULC change is changes in the Land Surface Temperature (LST) which has a major role in maintaining the local and regional climate patterns. Changes to the physical properties of the Earth’s surface are referred to as LULC changes. These include modifications to the plant cover, urbanization, agricultural methods, and natural disturbances like deforestation or afforestation. These adjustments immediately alter the surface energy distribution, which in turn affects the thermal characteristics of the land surface and its temperature [3, 4].

1.1 Land Surface Temperature The term Land Surface Temperature describes the temperature of the Earth’s surface, specifically the temperature of the soil, plants, rocks, bodies of water and other land surface phenomenon. LST is essential to many environmental processes and applications, including urban planning, agriculture, hydrology and climate modeling. Numerous variables such as solar radiation, air conditions, vegetation cover, surface characteristics and land use land cover have an impact on LST. It can alter both geographically and temporally reflecting changes in surface temperature throughout the day and across the seasons [5].

1.2 Urban Heat Island Urbanization also led to the formation of Urban Heat Island (UHI). Urban heat islands are formed when impermeable surfaces like buildings, roads, and pavements overtake naturally occurring landscapes, causing urban regions to experience greater LST than nearby rural or natural areas [6–8]. The local climate, human health, energy use, and general urban comfort may all be significantly impacted by its formation. On the other hand, modifications to the vegetation, including afforestation or deforestation, can also affect LST. Through mechanisms such as shade and evapotranspiration, vegetation has an impact on the surface energy balance. By limiting the quantity of solar energy received and encouraging water evaporation, which helps control temperature, vegetation cover has a cooling effect. As a result, changes in land cover that cause the loss or increase of vegetation can cause significant adjustments in LST patterns [9–11].

Effect of LULC Changes on Land Surface Temperature

157

1.3 Objectives Main aim of this paper is to analyze the Land Surface Temperatures of Hyderabad City • To Analyze the variations of Land Use Land Cover Changes over the past two decades. • To Estimate the change in surface temperatures due to LULC changes and to map the spatial patterns of LST. • To Predict the future change in LULC for Hyderabad city for the year 2027.

2 Study Area Hyderabad, a Metropolitan city in Telangana with a diverse and dynamic urban landscape is located between latitudes 17°18' N to 17°35' N and Longitudes 78°15' E to 78°40' E. The city originated on the banks of the river Musi. It is situated at an average elevation of 542 m above sea level and experiences a semi-arid type of climate. The study area for the current study includes in and around regions of Hyderabad till Outer Ring Road (ORR). Parts of Rangareddy, Sangareddy and MedchalMalkajgiri districts are covered inside of ORR. HMDA region extended till ORR has experienced rapid urbanization. The population of these combined regions in 2002 was approximately 5,800,000 and in 2022 was approximately 10,000,000 constituting an increase of 72% population. This rapid increase in population increased the percentage of concrete sand bituminous structures in the city resulting in greater loss of natural land cover in and around the city (Fig. 1).

2.1 Data Used In this present study, remotely sensed satellite images were used for the Land Use Land Cover classification and also for the generation and analysis of Land Surface Temperature (LST). For this Landsat 7 and Landsat 8 data have been used. Data sets for 2002, 2007 and 2012 have been used from Landsat 7 and for the years 2017 and 2022 Landsat 8 data is used (Table 1).

2.2 Methodology The Methodology for the proposed study has shown in Fig. 2.

158

R. Kummari et al.

Fig. 1 Study Area Map

Table 1 Data collection from satellite imagery Satellite

Type of sensor

Acquisition date

Path and row

Data source

Landsat 7

ETM+

07-04-2002

144, 048

USGS Earth Explorer

Landsat 7

ETM+

07-05-2007

144, 048

Landsat 7

ETM+

20-05-2012

144, 048

Landsat 8

OLI/TIRS

24-04-2017

144, 048

Landsat 9

OLI/TIRS

24-05-2022

144, 048

2.3 Image Pre-processing The Raw data obtained from LANDSAT contains different types of errors and for that image pre-processing is required to take over the geometric, atmospheric and radiometric correction. Cloud free data of the winter season was used as it doesn’t induce excess heat due to climate, only surface heat is dominant and would help in clearly analyzing the thermal properties and temperature changes.

Effect of LULC Changes on Land Surface Temperature

159

Fig. 2 Flow chart showing the methodology

The DN of images were converted into TO values which can be precisely used for calculating various indices error free. ERDAS imagine was used for the correction and the DOS 1 atmospheric correction was also done in QGIS Pre-processing. The first and vital step after data extraction helps in increasing the accuracy of the results.

2.4 Image Classification A Typical method of remote sensing analysis is Image Classification. The goal of image classification is to classify the image’s pixels into groups according to the types of land cover they represent by using a classifier algorithm, which groups pixels based on their reflectance value. Two types of classification exist in general, which are unsupervised and supervised classification. In Unsupervised classification, the algorithm is self-defined and the software itself performs the classification without human intervention. In Supervised classification, the analyst uses samples of pixels from the image to teach the algorithm to recognize spectral classifications. For each land cover class, adequate samples must be used. Each class thought to differ from the others spectrally. To create samples, a polygon comprised of representative pixels from a class is created, and those pixels are subsequently loaded. Using the data from pixels as a signature. Following the

160

R. Kummari et al.

collection of samples for each land cover class, the separability of the samples will be evaluated by looking at the signatures on a plot of wavelength versus reflectance value. Depending on the results, samples might need to be improved.

2.5 Spectral Indices 2.5.1

Normalized Difference Vegetation Index

Normalized Difference Vegetation Index (NDVI) is the mostly used Vegetation Index in the analysis of satellite images. It gives information about the density and health of vegetation cover based on the difference between the reflectance of Near-Infrared (NIR) and Red spectral bands. The health of plants depends on the electromagnetic spectrum. It is crucial to how NDVI functions and enables us to assess a plant’s health based on how it reflects light and energy. A plant appears green to the human eye because of the chlorophyll pigment, which reflects green wavelengths and absorbs red ones. Plants’ cell architecture reflects near-infrared (NIR) wavelengths. As a result, the plant develops, flourishes, and contains more cell structures when photosynthesis takes place. This implies that a healthy plant actively absorbs red light and reflects NIR—one with a lot of chlorophyll and cell structures. NDVI can be calculated using the below formula: NDVI =

NIR − RED NIR + RED

NDVI values range from −1 to +1 with −1 indicating no vegetation and +1 indicating dense healthy vegetation and forest cover [12, 13].

2.5.2

Normalized Difference Built-Up Index

Normalized Difference Built-up Index (NDBI) is the mostly used Built-up or urban index in the analysis of satellite images. It quantifies the difference in reflectance between the shortwave infrared (SWIR) and near-infrared (NIR) spectral bands. NDBI can be calculated using the below formula: NDVI =

SWIR − NIR SWIR + NIR

NDBI values range from −1 to +1. Greater positive values indicate a higher density of built-up areas. It highlights the contrast in reflectance between built-up structures (which have high reflectance in the SWIR band) and natural features like vegetation or bare soil (which have high reflectance in the NIR band).

Effect of LULC Changes on Land Surface Temperature

161

2.6 LST Estimation from Landsat 7 The bands that are used for LST estimation using Landsat 7 data is Band 6 (Low gain or High gain). Firstly, the bands are to be rectified for scan line errors. This can be done by imposing the gap mask layers. Deriving LST from satellite images follows 1. Conversion of DN to Radiance 2. Conversion of Radiance to Brightness Temperature (BT) 3. Conversion of Kelvin to Degree Celsius 2.6.1

Conversion of DN to Radiance

DN values of the satellite images have to be converted to spectral radiance using the equation Lλ =

L max − L min ∗ (QCal − QCalmin ) + L min QCalmax − QCalmin

where, Lλ QCal L max L min QCalmin QCalmax

2.6.2

Spectral radiance at the sensor (W/m2 *sr*µm). Quantized calibrated pixel value in DN. Spectral radiance scaled to QCalmax (W/m2 *sr*µm). Spectral radiance scaled to QCalmin (W/m2 *sr*µm). The minimum Quantized calibrated pixel value in DN. The maximum Quantized calibrated pixel value in DN.

Conversion of Radiance to Brightness Temperature (BT)

Here the converted spectral radiance of thermal band of ETM + sensor is further processed to temperature in Kelvin (K) by involving the following equation [12, 14–16]. BT =

ln

K2 ) ( K1 +1 Lλ

where, BT At satellite brightness temperature (Kelvin). Spectral radiance at the sensor (W/m2 *sr*µm). Lλ K1 K2 are calibration constants.

162

2.6.3

R. Kummari et al.

Conversion of Kelvin to Degree Celsius

Further, the converted temperature scale in K of the thermal band can be again transformed into degree Celsius (°C) using the below equation T (◦ C) = BT − 273.15

2.7 LST Estimation from Landsat 8 In this, Band 10 and Band 11 can be used for the generation of spectral radiance and brightness temperature. And NIR (Near-Infrared, band5) and Red (band4) bands are used for calculating Emissivity using NDVI method.

2.7.1

Conversion to TOA Radiance

Using the Radiance rescaling factor, Thermal Infra-Red Digital Numbers can be converted to TOA spectral radiance. Lλ = ML ∗ QCal + AL where, Lλ ML AL QCal

TOA Spectral radiance (W/m2 *sr*µm). Radiance multiplicative Band (No.) Radiance Add Band (No.) Quantized and Calibrated standard product pixel values (DN).

2.7.2

Top of Atmosphere (TOA) Brightness Temperature

Spectral radiance data can be converted to top of atmosphere brightness temperature using the thermal constant values in metadata file BT =

K2 ) − 273.15 ( K1 ln Lλ + 1

where, BT TOA brightness temperature (°C). Lλ Spectral radiance at the sensor (W/m2 *sr*µm). K1 and K2 are calibration constants.

Effect of LULC Changes on Land Surface Temperature

163

2.8 NDVI Method for Emissivity Correction Calculation 2.8.1

Calculate NDVI

NDVI can be calculated by using the following equation NDVI =

NIR − RED NIR + RED

where, NIR is Near-Infrared which is band 5 and red is band 4 [12, 13].

2.8.2

Proportion of Vegetation

Proportion of Vegetation can be obtained by using the max and min values of the NDVI [17] ( PV =

NDVI − NDVImin NDVImax − NDVImin

)2

where, PV NDVI NDVImin NDVImax 2.8.3

Proportion of Vegetation. DN values from NDVI image. Minimum DN values from NDVI image. Maximum DN values from NDVI image.

Land Surface Emissivity

Land Surface Emissivity (LSE) is the average emissivity of an element on the surface of the Earth calculated from the PV [18] E = 0.004 ∗ PV + 0.986 E Land surface emissivity. PV Proportion of Vegetation. 2.8.4

Land Surface Temperature

Land Surface Temperature is the radiative temperature which is calculated using Top of Atmosphere brightness temperature, wavelength of emitted radiance and Land surface emissivity [19–21]

164

R. Kummari et al.

( LST =

BT

(

( ( ) ) 1 + 0.00115∗BT ∗ ln(E) 1.4388

where, BT Top of atmosphere brightness temperature (°C). E Land Surface Emissivity.

2.9 Future Prediction of LULC The Land use Land cover is the most important parameter for the LST analysis. The past LULC change dynamics would help in prediction the of Future prediction. Various methods like Random Forest, Artificial Neural Network (ANN), CA and multiple regression QGIS software were used for the LULC prediction over Hyderabad city and surrounding area up to ORR. The MOLUSCE (Modules for Land Use Change Evaluation) plugin tool was used for LULC simulation and prediction. The known algorithms CA, ANN, Logistic Regression and MCDA (Multiple criteria decision analysis) were used for the study area. The LULC data is validated with the predicted LULC map and later the future prediction is performed. Pearson Correlation Constant was generated first and the area changes for different years and a transition matrix was generated. Transition potential modeling was used, the neural network was trained and the graph was generated between the trained data and the input data. Then simulation was under these spatial variables and after that validation was done for predicted 2022 and actual 2022. Using this relation, the future LULC for the years 2032 and 2042 are to be estimated. The model that validates the simulated LULC map generates Kappa characteristics such as Overall kappa, Kappa histogram, Kappa location and % of correctness which are used for evaluating the accuracy of the model are given below, K =

P(A) − P(E) 1 − P(E)

A)−P(E) . Kappa location: K loc = P( Pmax −P(E) Pmax −P(E) Kappa Histogram: K h = 1−P(E) . Σc Σc = = = i=1 pii i=1 pi T pT i , Pmax ΣcWhere, P(A) = P( A) min( p p is i, jth cell of contingency table, p is the sum of all p ), i T T i i j i T l=i cells in the ith row, pT i is the sum of all cells in the jth column and c is the raster categories count.

Effect of LULC Changes on Land Surface Temperature

165

3 Results and Discussions 3.1 Land Use Land Cover Analysis In the present study, four land use land cover classes have been classified such as Vegetation, Waterbodies, Built-up and Barren Lands. It was discovered that the study area, which includes Hyderabad and its environs, had a distinct LC with the urban class dominating in the center. Urban class was confined to the study area’s center part in 2001, when arid plains dominated the environment. And some parts of the north parts of the city were lined with dense vegetation, and there were also sections of it on the edges of the city. Its differentiation of the Land use for the years 2002, 2007, 2012, 2017 and 2022 can be viewed from the Land use land cover images (Fig. 3) and the area and percentage changes are listed in Table 2 and Fig. 3. By analyzing the data from Table 2, there has been a significant decrease in barren lands over the years. In 2002, the Urban class covered an area of 307.23 km2 , which grew to 538.76 km2 by 2022, indicating a substantial percentage increase of 75.36% over the two-decade period. Barren land, on the other hand, has declined by 49.6% during the same timeframe. Vegetation has exhibited an upward trend, particularly from 2012 onwards. There was a 19.99% increase from 2012 to 2017, followed by a further 7.14% increase from 2017 to 2022. Overall, the vegetation area expanded from 382.8 to 480.32 km2 over the two-decade period. Furthermore, water bodies have shown an overall increase of 11% in the past two decades. These findings highlight the dynamic changes in land use and land cover in Hyderabad, with significant implications for urban development and environmental management.

3.2 NDVI The Normalized Difference Vegetation Index (NDVI) is a spectral index used to represent the concentration of vegetation. In this study, NDVI values for the years 2002, 2007, 2012, 2017, and 2022 were generated using Band 4 (Red) and Band 5 (NIR) imagery. NDVI values ranging from −1 to +1 indicate the presence and density of vegetation. Positive values suggest areas with thick and dense vegetation, while negative values indicate sparse or no vegetation. In the study area, NDVI values range between −0.49 and +0.53, revealing a varied pattern of vegetation coverage. It is evident that certain parts of the city lack significant vegetation (Fig. 4).

319.36

647.16

307.23

689.91

Built-up

Barren

389.17

km2

629.23

367.25

128.32

373.61

2012

Total extent of study area: 1498.41

142.72

382.8

118.47

Vegetation

2007

2002

Area (km2 )

Waterbodies

Land cover

445.37

483.22

121.51

448.31

2017

347.71

538.76

131.62

480.32

2022

−6.2

3.95

20.47

1.66

2002–2007

−2.77

15 −29.22

31.58

19.99 −5.31

−4

2012–2017

−10.09

2007–2012

Change for each 5 years

Percentage changes for each class

Table 2 Area and percentage change over study area for different years

−21.92

11.49

8.32

7.14

2017–2022

10 years

−8.8

19.54

8.31

−2.4

2002–2012

−44.74

46.7

2.57

28.56

2012–2022

20 years

−49.6

75.36

11.1

25.48

2002–2022

166 R. Kummari et al.

Effect of LULC Changes on Land Surface Temperature

167

Fig. 3 LULC images

3.3 Land Surface Temperature Analysis Land surface temperature is generated by using the spectral indices such as radiance, brightness temperature, proportion of vegetation and land surface emissivity. Hyderabad city is showing a mixed type of climate. The outer parts of the city excluding the central part have been experiencing greater temperatures. In the year 2002, only certain parts of the city experienced a temperature more than 35 °C and the max temperature was recorded as 42 °C. But for the year 2012, Hyderabad city has experienced a heat wave. Due to the lack of vegetation, the temperature has reached a maximum of 57.59 °C. And from the year 2017, the percentage of vegetation in and

168

R. Kummari et al.

Fig. 4 NDVI images

around the parts of the city have been increasing resulting in a decrease in temperatures when compared with 2012. The maximum temperature recorded in 2017 was 50.2 °C and that of in 2022 was 41 °C. LULC have made a significant impact on the temperatures for the city as the vegetation percentage has increased over the years (Table 3 and Fig. 5).

Effect of LULC Changes on Land Surface Temperature Table 3 Max and min land surface temperatures

169

18.15

Max temp

Min temp

2002

42

21.41

Mean temp 31.705

2007

35.92

18.82

27.37

2012

57.69

18.15

37.92

2017

50.52

20.3

35.41

2022

40.55

19.85

30.2

Fig. 5 Image showing LST variations

170

R. Kummari et al.

Table 4 Pearson’s correlation

Initial

Final

Pearson’s correlation

2002

2007

0.7912

2007

2012

0.8123

2012

2017

0.7879

2017

2022

0.76215

3.4 Future Prediction 3.4.1

Correlation Evaluation

Pearson correlation coefficient is the measure of the strength of linear relation between two variables. The value of this coefficient ranges between −1 to +1. Negative value indicates no correlation and positive values indicate a stronger linear relation. From the result, it is viewed that the years between 2002 and 2022 are more strongly related to linear correlation (Table 4).

3.4.2

Validation of the Model

The predicted LULC is to be validated with the existing datasets for its reliability. For this validate option in Molusce plugin is used. Two data sets are used for validation. The module generates different kappa coefficients such as kappa histogram, kappa overall, kappa location and % of correctness. Initial input data set is given the 2002 LULC and the final dataset is given the 2007 LULC map for the initial simulation of LULC for the year 2012. Kappa statistics with 0% indicates that there is no agreement and 100% indicates perfect agreement. Table 5 shows the result of the validated module to check the agreement between the simulated 2012 raster and the real classified raster of 2012. The estimation of kappa (overall) is 0.7113, kappa (histogram) is 0.8127 and kappa (location) is 0.7095 while the % of correctness is 72.7421, which shows the consistency between the predicted 2012 LULC and the real 2012 LULC is good. And similarly, the model is used for simulating LULC for the years 2017 and 2022. The simulated maps are then validated with the reference 2017 and 2022 rasters respectively. On observing the % correction and other kappa indices, the model is Table 5 Kappa index generated from model validation

2002_07_12

2002_07_17

2002_07_22

% of correctness

72.7421

71.2379

69.9787

Kappa histogram

0.8127

0.7911

0.8871

Kappa overall

0.7113

0.7593

0.7261

Kappa location

0.7095

0.7293

0.7302

Effect of LULC Changes on Land Surface Temperature

171

Fig. 6 Simulated LULC maps for the years 2012, 2017 and 2022

showing consistency for the simulated and real LULC. So this model suits best for Hyderabad city for the future prediction of LULC for the year 2027 (Fig. 6). Using these inputs, the LULC for the year 2027 is predicted. From Fig. 7, it can be viewed that the built-up in the study area has been increased to a great extent by a decrease in the barren lands.

4 Conclusions By using data from Landsat 7 ETM+ and Landsat 8 OLI/TIRS, this study has investigated about the Land Use Land Cover (LULC) changes over the past two decades, specifically from 2002 to 2022. It has also made an impact in studying about the Land Surface Temperature for the Hyderabad City with an extension up to Outer Ring Road (ORR). During this study, it is observed that Barren lands have experienced a major decrease of 49.6% from 2002 to 2022. Vegetation over the study area has witnessed the least increase in the period 2002–2012. But from 2012, there has

172

R. Kummari et al.

Fig. 7 LULC map predicted for the year 2027

been a better development of vegetation. There has been an increase of 25.48% from 2012 to 2022. Notably, the built-up in the city region has a growth in each period. It has experienced an overall increase of 75.36% from 2002 to 2022. Water bodies have experienced a least increase of 11.1% over the last two decades. The surface temperature in the study region was evaluated using spectral indices such as radiance, brightness temperature, NDVI, proportion of vegetation, and emissivity. Land surface temperatures exhibited different trends over the past years. The lowest temperature recorded was 18.15 °C, while the maximum surface temperature of 57.69 °C was observed in 2012. Between 2012 and 2017, the city experienced higher temperatures, reaching 50 °C and above. However, due to an increase in vegetation from 2017 onwards, there has been a slight decrease in temperatures. The aim of this research is to predict future Land Use and Land Cover (LULC) using QGIS and the Molusce plugin, employing a cellular automata-Markov chainbased simulation. LULC rasters were simulated for the years 2012, 2017, and 2022, which were then validated against reference LULC maps. The kappa indices indicate that the model provides the best fit for predicting LULC changes in the study area. A LULC map for the year 2027 was generated, revealing a significant increase in the built-up class and a considerable decrease in barren lands. These findings contribute to a better understanding of the LULC dynamics and highlight the changing patterns of land cover in the study area. Further analysis and

Effect of LULC Changes on Land Surface Temperature

173

interpretation of these results can aid in urban planning, natural resource management, and decision-making processes to ensure sustainable development and efficient land use practices.

References 1. Dwivedi A, Khire MV (2018) Application of split-window algorithm to study urban heat island effect in Mumbai through land surface temperature approach. Sustain Cities Soc 41:865–877. https://doi.org/10.1016/j.scs.2018.02.030 2. John J, Bindu G, Srimuruganandam B, Wadhwa A, Rajan P (2020) Land use/land cover and land surface temperature analysis in Wayanad district, India, using satellite imagery. Ann GIS 26(4). https://doi.org/10.1080/19475683.2020.1733662 3. Singh P, Kikon N, Verma P (2017) Impact of land use change and urbanization on urban heat island in Lucknow city, Central India. A remote sensing-based estimate. Sustain Cities Soc 32:100–114. https://doi.org/10.1016/j.scs.2017.02.018 4. Sharma R, Pradhan L, Kumari M, Bhattacharya P (2021) Assessing urban heat islands and thermal comfort in Noida City using geospatial technology. Urban Clim 35:100751. https:// doi.org/10.1016/j.uclim.2020.100751 5. Buscail C, Upegui E, Viel J-F (2012) Mapping heatwave health risk at the community level for public health action. Int J Health Geogr 11:38. http://www.ij-healthgeographics.com/content/ 11/1/38 6. Halder B, Bandyopadhyay J, Banik P (2021) Monitoring the effect of urban development on urban heat island based on remote sensing and geo-spatial approach in Kolkata and adjacent areas, India. Sustain Cities Soc 74:103186. https://doi.org/10.1016/j.scs.2021.103186 7. El-Hattab M, Amany SM, Lamia GE (2017) Monitoring and assessment of urban heat islands over the Southern region of Cairo Governorate, Egypt. Egypt J Remote Sens Space Sci 21:311– 323. https://doi.org/10.1016/j.ejrs.2017.08.008 8. Kamboj S, Ali S (2020) Urban sprawl of Kota city: a case study of urban heat island linked with electric consumption. Mater Today: Proc 46(11). https://doi.org/10.1016/j.matpr.2020.08.783 9. Deilami K, Kamruzzaman M, Liu Y (2018) Urban heat island effect: a systematic review of spatio-temporal factors, data, methods, and mitigation measures. Int J Appl Earth Obs Geoinformation 67:30–42. https://doi.org/10.1016/j.jag.2017.12.009 10. Monteiro FF, Gonçalves W, Melo LD, Andrade B (2021) Assessment of urban heat islands in Brazil based on MODI’S remote sensing data. Urban Clim 35:100726. https://doi.org/10. 1016/j.uclim.2020.100726 11. Equere V, Mirzaei PA, Riffat S, Wang Y (2021) Integration of topological aspect of city terrains to predict the spatial distribution of urban heat island using GIS and ANN. Sustain Cities Soc 69:10285. https://doi.org/10.1016/j.scs.2021.102825 12. Aniello C, Morgan K, Busbey A, Newland L (1995) Micro-urban heat islands using Landsat TM and a GIS. Comput Geosci 21(8):965–969. https://doi.org/10.1016/0098-3004(95)00033-5 13. Estoque RC, Murayama Y, Myint SW (2017) Effects of landscape composition and pattern on land surface temperature: an urban heat island study in the megacities of Southeast Asia. Sci Total Environ 577:349–359. https://doi.org/10.1016/j.scitotenv.2016.10.195 14. Mohan M, Kandya A, Battiprolu A (2011) Urban heat island effect over national capital region of India: a study using the temperature trends. J Environ Prot 2:465–472. https://doi.org/10. 4236/jep.2011.24054 15. Shahmohamadi P, Che-Ani AI, Ramly A, Maulud KNA, Mohd-Nor MFI (2010) Reducing urban heat island effects: a systematic review to achieve energy consumption balance. Int J Phys Sci 5(14):2202–2207. https://doi.org/10.5897/IJPS.9000475

174

R. Kummari et al.

16. Weng Q, Lub D, Schubring J (2004) Estimation of land surface temperature-vegetation abundance relationship for urban heat island studies. Remote Sens Environ 89(4):467–483. https:// doi.org/10.1016/j.rse.2003.11.005 17. Valor E, Caselles V (1996) Mapping land surface emissivity from NDVI: application to European, African and South American areas. Remote Sens Environ 57:167–184. https://doi.org/ 10.1016/0034-4257(96)00039-9 18. Van de Griend A, Owe M (1993) On the relationship between thermal emissivity and the normalized difference vegetation index for natural surfaces. Int J Remote Sens 14:1119–1131. https://doi.org/10.1080/01431169308904400 19. Avdan U, Jovanovska G (2016) Algorithm for automated mapping of land surface temperature using LANDSAT 8 Satellite Data. https://doi.org/10.1155/2016/1480307 20. Roy DP, Wulder MA, Loveland TR, Woodcock CE, Allen RG, Anderson MC, Helder D, Irons JR, Johnson DM, Kennedy R, Scambos TA, Schaaf CB, Schott JR, Sheng Y, Vermote EF, Belward AS, Bindschadler R, Cohen WB, Gao F, Hipple JD, Hostert P, Huntington J, Justice CO, Kilic A, Kovalskyy V, Lee ZP, Lymburner L, Masek JG, McCorkel J, Shuai Y, Trezza R, Vogelmann J, Wynne RH, Zhu Z (2014) Landsat-8: science and product vision for terrestrial global change research. Remote Sens Environ 145:154–172. https://doi.org/10.1016/j.rse.2014. 02.001 21. Yu X, Guo X, Wu Z (2014) Land surface temperature retrieval from Landsat 8 TIRS—comparison between radiative transfer equation-based method, split window algorithm and single channel method. Remote Sens 6:9829–9852. https://doi.org/10.3390/rs6109829

Estimation of Aerosol Direct Radiative Forcing in Southern India K. Tharani, Deva Pratap, Keesara Venkatareddy, and P. Teja Abhilash

Abstract The Aerosol Direct Radiative Forcing is a prominent parameter which is used to assess the effect of aerosols on temperature. Previous studies have shown the calculation of ADRF under clear sky conditions and model simulations to know the impact on changes in temperature. In the current study, a novel method is implemented to estimate ADRF under all-sky conditions using Modern-Era Retrospective analysis for Research and Applications (MERRA-2) data. The radiation diagnostics of MERRA-2 provide downward short wave and long wave fluxes under all-sky conditions but not the upwelling fraction of longwave radiation after interacting with the atmosphere. The upwelling fraction is estimated by factored method and difference method in this study and the time series is compared with the available downward flux data. The upwelling longwave flux data from the factored method is found to be appropriate in terms of matching the time series when compared to the data from the difference method. The estimation is carried out for five grid points, each from five climate zones of the study area and the times series plots have shown a better insight on the implemented factored method for the estimation of longwave flux. The ADRF is calculated using the available shortwave fluxes and estimated longwave fluxes at three different levels of atmosphere namely Top of Atmosphere (TOA), Surface and within the atmosphere for the year 2019. The ADRF at TOA and at surface level show that the warm semi-arid and sub-tropical oceanic highland climate regions have experienced higher ADRF when compared to the remaining climate zones. The ADRF in the atmosphere shows that positive ADRF was seen in the same climate regions. This illustrates that the rate of heating is more in the warm semi-arid and sub-tropical oceanic highland climate region in the year 2019. Keywords Aerosol direct radiative forcing · Shortwave flux · Upwelling longwave flux

K. Tharani (B) · D. Pratap · K. Venkatareddy Research Scholar, Department of Civil Engineering, NIT Warangal, Warangal, India e-mail: [email protected] P. T. Abhilash Andhra University, Visakhapatnam, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_13

175

176

K. Tharani et al.

1 Introduction The major socio-economic issue in the current scenario is global and regional climate variability, which is due to an increase in urban emission concentrations. It affects the radiation budget of earth’s surface and thus influences the climate parameters to a considerable extent. One of the components of urban emissions is aerosols. They are the minute particles suspended in the air or in any other gas. Aerosols have an impact on the climate system and weather in two primary ways. The first is through scattering and absorption of radiation that is considered as the direct effect of aerosols, which consequently, modifies the Earth’s radiation budget [1, 3, 5, 7]. The absorbing aerosols (black carbon, mineral dust, etc.) decrease the Earth’s albedo and heat the atmosphere which results in the greenhouse effect near the Earth’s surface [13–15], whereas scattering aerosols (sulfate, nitrate, etc.) will increase the Earth’s albedo and decrease the solar radiation reaching the Earth which results in a net cooling of the Earth-atmosphere system [2, 4, 6, 17]. Secondly, aerosols can alter cloud microphysical properties, increase cloud lifetime and enhance precipitation efficiency [9, 11, 16, 18]. Aerosol studies over Peninsular India focus primarily on aerosol loading, types and properties of aerosols and influence on the climate [10, 12]. The effect of aerosols on consecutive warming and cooling of the atmosphere is studied by using Aerosol Direct Radiative Forcing (ADRF). The main objective of the present study is to estimate ADRF under all-sky conditions.

2 Data The present study concentrates on the direct effect of aerosols on the Earth’s radiation budget. In order to understand the process, it is essential to have an insight on the forcing component of aerosols. Generally, radiative forcing is used to determine the effect of a particular component in the considered phenomenon. Hence, Aerosol Direct Radiative Forcing (ADRF) is defined as the component which describes the effect of aerosols on the radiative fluxes at top of atmosphere (TOA), at surface and absorption of radiation within the atmosphere. The study area is confined to peninsular India comprising of 9 states, which are divided into five climate zones as shown in Fig. 1. The datasets used are NetCDF files of Modern-Era Retrospective Analysis for Research and Applications (MERRA-2). The parameters from radiation diagnostics of MERRA-2 (M2T1NXRAD) are used to calculate ADRF at TOA, surface and in the atmosphere.

Estimation of Aerosol Direct Radiative Forcing in Southern India

177

Fig. 1 Study area

3 Methodology The methodology followed for the current study is shown in Fig. 2. Firstly, the parameters in NetCDF format are collected from MERRA-2 using GES DISK (https:// disc.gsfc.nasa.gov/) platform. Then the parameters are averaged daily and written down in Excel using R statistical software. The missing parameters namely Surface Net Downward Longwave Flux assuming no aerosol (LWGNTCLN) and Upwelling Longwave Flux at TOA assuming no aerosol (LWTUPCLN) are calculated by the factorial method as given in Eq. 3. Secondly, ADRF at TOA and surface is calculated using the Eqs. 1 and 2 respectively. Finally, ADRF in the atmosphere (ADRF ATM ) is calculated by the difference of ADRF at TOA and at the surface.

178

K. Tharani et al.

Fig. 2 Methodology

3.1 Calculation of ADRF The parameters required for the calculation of ADRF are Surface Net Downward Shortwave Flux (SWGNT), Surface Net Downward Shortwave Flux assuming no aerosol (SWGNTCLN), Surface Net Downward Longwave Flux (LWGNT), Surface Net Downward Longwave Flux assuming no aerosol (LWGNTCLN), TOA Net Downward Shortwave Flux (SWTNT), TOA Net Downward Shortwave Flux assuming no aerosol (SWTNTCLN), Upwelling Longwave Flux at TOA (LWTUP), Upwelling Longwave Flux at TOA assuming no aerosol (LWTUPCLN) [8]. Out of all these parameters, six are directly available from MERRA-2. The LWGNTCLN and LWTUPCLN are not available from MERRA-2 which need to be calculated. Generally, the Aerosol Direct Radiative Forcing at surface Eq. (1) and at TOA Eq. (2) are calculated as follows: AD R FSU R = (SW G N T + L W G N T ) − (SW G N T C L N + L W G N T C L N ) (1) AD R FT O A = (SW T N T + L W T U P) − (SW T N T C L N + L W T U PC L N ) (2) The normal convention says that positive ADRF ATM warms up the atmosphere and vice versa. The current study aims to obtain the missing parameters by two methods namely factored method and difference method. The factored method uses the factor

Estimation of Aerosol Direct Radiative Forcing in Southern India

179

obtained by dividing SWGNTCLN with SWGNT and uses the value to calculate LWGNTCLN as shown in Eq. (3). The difference method uses the difference between SWGNTCLN with SWGNT and uses the value to calculate LWGNTCLN as shown in Eq. (4). LW GN T C L N =

SW G N T C L N ∗ LW GN T SW G N T

L W G N T C L N = SW G N T C L N − SW G N T + L W G N T

(3) (4)

As a part of the analysis, five grid points each in a climate zone are considered and a time series plot for the data corresponding to 2019 of long wave flux calculated by factored and difference method is plotted to match the peaks and dips. The five grid points are Goa (Am), Nanded (Aw), Hyderabad (BSh), Raipur (Cwa) and Coimbatore (Cwb).

4 Results and Discussions The time series plot for the calculated LWGNTCLN using factored and difference methods along with the available LWGNT are shown in Fig. 3. The parameter calculated by the difference method shows disparity in the values at peaks and dips and that calculated by the factored method matches with the existing data without aerosol. However, the occurrence of peaks and dips is same for both the methods. Hence, the data calculated by factored method is chosen to be the suitable one as it is in accordance with the parameter without aerosol.

4.1 ADRF at TOA, Surface and in the Atmosphere The yearlong ADRF for the Peninsular India is calculated for the year 2019 at three different levels (Fig. 4). Most of the warm semi-arid, tropical monsoon, sub-tropical oceanic highland climate regions have witnessed a higher range of ADRF at TOA in the study area. The tropical savanna region has seen a lower ADRF at TOA in the year 2019. A similar pattern was visible at the surface except that a part of the tropical savanna towards the East coastal line has seen a reduction in ADRF. The results show that the ADRF is high towards the southern part of the study area, moderate in the central region and low towards the northern part of the study area. Almost 159 lakh ha of the study area has experienced a positive ADRF whereas the remaining area has experienced negative ADRF in the atmosphere. Generally, positive ADRF indicates that there is a rise in temperature and vice versa. The positive ADRF in the range of 0–1 is distributed over warm semi-arid and sub-tropical oceanic highland climate regions. Further correlation with air temperature is required to assess

180

K. Tharani et al.

Fig. 3 Time series plot of LWGNT, LWGNTCLN_f, LWGNTCLN_d for a Goa b Nanded c Hyderabad d Raipur e Coimbatore for the year 2019

a)

b)

c)

d)

e)

the effect of aerosols in the atmosphere. Also, the heating rate in each climate zone can help us to know the changes in global warming.

Estimation of Aerosol Direct Radiative Forcing in Southern India

181

koppen_south

ADRF -20 - -15 -15 - -10 -10 - -8

a)

-8 - -6

b)

-6 - -4 -4 - -2 -2 - 0

koppen_south

ADRF_atm -5 - -4 -4 - -3 -3 - -2 -2 - -1 -1 - 0

c)

0-1

Fig. 4 ADRF at a Top of atmosphere b Surface c In the atmosphere

5 Conclusions The current study has estimated the ADRF under all-sky conditions using MERRA2 data. The upwelling longwave fluxes which were not available in the MERRA-2 radiation diagnostics data were estimated by factored and difference methods. The following conclusions are drawn from the analysis performed. • The upwelling longwave flux estimated by the difference method was found to be inconsistent when compared to the available flux data. However, the peaks and dips of the time series curve were found to be in accordance. • The upwelling longwave flux estimated by factored method was in agreement with the available flux data. Hence, the factored method was found to be the best one for the estimation of flux data.

182

K. Tharani et al.

• The estimated and available fluxes were used to calculate radiative forcing by aerosols. The ADRF was calculated at TOA, surface and within the atmosphere for the year 2019. • The ADRF at TOA and surface was found to be negative in the study area whereas the ADRF within the atmosphere was positive only in the sub-tropical oceanic highland climate and a part of warm semi-arid climate regions. • The positive ADRF might be attributed to the entrapment of radiation in the lower layers of the atmosphere, which might lead to atmospheric warming.

References 1. Adesina AJ, Kumar KR, Sivakumar V, Griffith D (2014) Direct radiative forcing of urban aerosols over Pretoria (25.75 S, 28.28 E) using AERONET sunphotometer data: first scientific results and environmental impact. J Environ Sci 26(12):2459–2474 2. Alam K, Trautmann T, Blaschke T (2011) Aerosol optical properties and radiative forcing over mega-city Karachi. Atmos Res 101(3):773–782 3. Aruna K, Kumar TL, Murthy BK, Babu SS, Ratnam MV, Rao DN (2016) Short wave aerosol radiative forcing estimates over a semi urban coastal environment in south-east India and validation with surface flux measurements. Atmos Environ 125:418–428 4. Dhar P, De BK, Banik T, Gogoi MM, Babu SS, Guha A (2017) Atmospheric aerosol radiative forcing over a semi-continental location Tripura in North-East India: model results and ground observations. Sci Total Environ 580:499–508 5. Feng H, Zou B (2019) Satellite-based estimation of the aerosol forcing contribution to the global land surface temperature in the recent decade. Remote Sens Environ 232:111299 6. García OE, Díaz JP, Expósito FJ, Díaz AM, Dubovik O, Derimian Y, Dubuisson P, Roger JC (2012) Shortwave radiative forcing and efficiency of key aerosol types using AERONET data. Atmos Chem Phys 12(11):5129–5145 7. Guleria RP, Kuniyal JC (2016) Characteristics of atmospheric aerosol particles and their role in aerosol radiative forcing over the northwestern Indian Himalaya in particular and over India in general. Air Qual Atmos Health 9(7):795–808 8. https://disc.gsfc.nasa.gov/ 9. Jing X, Suzuki K, Michibata T (2019) The key role of warm rain parameterization in determining the aerosol indirect effect in a global climate model. J Clim 32(14):4409–4430 10. Johnson JS, Regayre LA, Yoshioka M, Pringle KJ, Lee LA, Sexton DM, Rostron JW, Booth BB, Carslaw KS (2018) The importance of comprehensive parameter sampling and multiple observations for robust constraint of aerosol radiative forcing. Atmos Chem Phys 18(17):13031–13053 11. Kant S, Panda J, Pani SK, Wang PK (2019) Long-term study of aerosol–cloud–precipitation interaction over the eastern part of India using satellite observations during pre-monsoon season. Theoret Appl Climatol 136(1):605–626 12. Kaskaoutis DG, Sinha PR, Vinoj V, Kosmopoulos PG, Tripathi SN, Misra A, Sharma M, Singh RP (2013) Aerosol properties and radiative forcing over Kanpur during severe aerosol loading conditions. Atmos Environ 79:7–19 13. Kumar KR, Kang N, Sivakumar V, Griffith D (2017) Temporal characteristics of columnar aerosol optical properties and radiative forcing (2011–2015) measured at AERONET’s Pretoria_CSIR_DPSS site in South Africa. Atmos Environ 165:274–289 14. Lu Q, Liu C, Zhao D, Zeng C, Li J, Lu C, Wang J, Zhu B (2020) Atmospheric heating rate due to black carbon aerosols: Uncertainties and impact factors. Atmos Res 240:104891

Estimation of Aerosol Direct Radiative Forcing in Southern India

183

15. Patel PN, Dumka UC, Kaskaoutis DG, Babu KN, Mathur AK (2017) Optical and radiative properties of aerosols over Desalpar, a remote site in western India: source identification, modification processes and aerosol type discrimination. Sci Total Environ 575:612–627 16. Qian YUN, Leung LR, Ghan SJ, Giorgi F (2011) Regional climate effects of aerosols over China: modeling and observation. Tellus B: Chem Phys Meteorol 55(4):914–934 17. Ramachandran S, Rupakheti M, Lawrence MG (2020) Aerosol-induced atmospheric heating rate decreases over South and East Asia as a result of changing content and composition. Sci Rep 10(1):1–17 18. Yu H, Dickinson RE, Chin M, Kaufman YJ, Zhou M, Zhou L, Tian Y, Dubovik O, Holben BN (2004) Direct radiative effect of aerosols as determined from a combination of MODIS retrievals and GOCART simulations. J Geophys Res: Atmos 109(D3)

Estimation of Groundwater Potential Zones in Southern Dry Agro-Climatic Area Using Geoinformatics and AHP Technique A. B. Gireesh and M. C. Chandan

Abstract The world’s demand for groundwater has been severely strained by overuse of groundwater and major climatic change over time. As the global need for drinking water for human consumption, agriculture, and industrial applications grows, so does necessity to assess groundwater. Due to the quick access to data, analysis, and knowledge, they give about resource for further improvement, GISbased studies have grown in importance in groundwater research in recent years. In order to identify the groundwater potential zone of southern dry agro-climatic area of Mysuru and Mandya District, India, the current study has been carried out. A total of 15 theme layers were established and researched to help define groundwater potential zones. Depending on respective attributes and water potential capabilities, Analytic Hierarchy Process determines weights allocated to each class in all thematic layers. Utilizing data on groundwater prospects in area (CGWB), study’s output was cross-validated, and total accuracy of methodology was 80.4%. The resulting groundwater potential zone was divided into three classifications: high, moderate, and low. According to the research, a moderate groundwater potential zone encompasses 65.8% of study area. There are zones with low and high groundwater potential in 6.75% and 27.43% of the area, respectively. The R2 value of 0.8 further demonstrates that estimated groundwater potential index and groundwater level values in recommended AHP model are quite reliable in predicting the outcome. The demarcation of groundwater potential zones has been highlighted as a critical step towards achieving Sustainable Development Goals (SDG 6 and 13), contributing to the implementation of sustainable water and land management. Keywords Groundwater potential zone · Geoinformatics · Analytical hierarchical process (AHP) · Sustainable Development Goals

A. B. Gireesh · M. C. Chandan (B) Department of Civil Engineering, The National Institute of Engineering, Mysuru 570008, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_14

185

186

A. B. Gireesh and M. C. Chandan

1 Introduction In addition to contributing to an area’s economic growth and biodiversity, groundwater is one of the most valued sources of water for agricultural, urban, and industrial activities worldwide [16]. A significant source of freshwater, groundwater presently provides around 34% of the total yearly water supply. Therefore, evaluating this resource is essential for the long-term sustainability of groundwater management systems [17]. Water stress is regarded as one of the most crucial socio-environmental issues across several countries because of groundwater demand is increasing, and overexploitation of this vital resource is threatening future generations [24]. Increased groundwater withdrawal has manifested in a gradual decline in the water table. Groundwater occurrence and distribution are influenced by a variety of natural and anthropogenic causes such as topographical, hydrological, biological, geological, and atmospheric influences [6, 22, 27, 32, 34]. Novel strategies are required to produce accurate and helpful information for decision-makers since methods for determining the factors that affect the spatial distribution of groundwater and approaches for acquiring data are developing [1]. The first and most critical stage in groundwater management is groundwater potential mapping (GPM) [29]. Groundwater potential mapping has been described as a strategy for systematic water resource development and planning [9]. Accordingly, for the judicious use of water resources, a groundwater management strategy is recommended. The primary focus of conventional techniques of identifying, characterizing, and mapping groundwater potential zones is ground surveys employing geophysical, geological, and hydrogeological equipment and a majority of which is expensive and time-consuming [5, 15, 24, 33]. Computer technology has been increasingly used in hydrological research in recent decades. The analysis, forecasting, and management of groundwater resources have all made substantial use of remote sensing (RS) and geographic information system (GIS) techniques [2, 8, 14]. Numerous research has been conducted using various multi-criteria decisionmaking approaches, geophysical methods, machine learning algorithms, and hybrid artificial intelligence. Analytical hierarchy process [3, 7, 15, 29], weight of evidence (WOE) [11, 34], frequency ratio (FR) [13, 26, 27], electrical resistivity, and vertical electrical sounding (VES) [18] (Fitteman and Stewart 1986) are some of the methods used by the researchers. The goal of current research is to estimate groundwater potential zones using geoinformatic and AHP method. The study objectives are to: (a) mapping groundwater zones using GIS methods; (b) establishing a robust model for estimating groundwater prospective zones, which will aid in agriculture production using GW and; (c) validate the obtained results with field data. As a result, findings of study provide vital recommendations and implementation measures to be taken by decision-makers, administrators, policymakers, and academics.

Estimation of Groundwater Potential Zones in Southern Dry …

187

2 Materials and Method The research area comprises Mysuru and Mandya district, which are situated in India’s southern region (Fig. 1). The geographical extents of study area are latitude of 11°73’N to 13°04’N and Longitude 75°90’E to 77°32’E, which consist of an area of 11,815 km2 . It is located within the Kaveri River’s watershed, which runs through northwestern and eastern parts of area. The Mysuru district’s soil types are divided into three categories as red sandy, red loamy, and deep black soils. Except for a small portion of T. Narasipur taluk, almost the entire district is covered in red sandy soil. The Mandya District’s soils range from red sandy loams to red clay loams, which are thin in hills and higher elevations but thick in valleys. The topographical elevation of study area ranges from 392 to 1175 m above mean sea level. According to the Koppen climatic classification, Mysore and Mandya have a tropical savanna climate that borders on a hot semi-arid climate. The district has moderate summer temperatures, up to 35 °C, and moderate winter temperatures, down to 20 °C. The average annual rainfall of study area ranges from 650 to 700 mm. In this study, 178 observation wells are considered monitored by Central Ground Water Board (CGWB). The groundwater level fluctuates from 0.4 to 19.6 m/bgl, according to CGWB data. The study area majorly consists of both urban as well as agricultural irrigated land. The behavior and flow of groundwater sources are influenced by a number of variables. Their impact varies by region, therefore identifying the proper one is critical for determining the possible groundwater zone. Literature suggest the goal of

Fig. 1 Map of research area

188

A. B. Gireesh and M. C. Chandan

the research, the nature of the region, and their availability influence their selection for investigation [21, 35]. The MCDM technique, i.e., the AHP utilized in this research for groundwater potential zone delineation, can be simplified down into five steps: thematic layer processing; pair-wise comparative matrix; weight calculation; consistency ratio; GWPZ map and validation.

2.1 Thematic Layer Processing A total of 15 layers of information were collected for this study’s groundwater potential zone delineation, including Geology, Geomorphology, Slope, Aspect, Elevation, Rainfall, Lineament Density, LULC, NDVI, Profile Curvature, Plan Curvature, Soil, TWI, Stream Power Index, and Slope length. Table 1 shows data used for estimating groundwater potential.

2.2 Analytic Hierarchy Process In 1980, Saaty brought out the Analytic Hierarchy Process (AHP) approach. AHP is a problem-solving method that divides issues into groups and then arranges them in a hierarchical framework. To select priority criteria, this technique combines a comparison of criteria with a pre-determined measuring scale. The major input of the AHP technique is expert perception; hence, subjectivity plays a role in retrieval decisions (Putra et al. 2018). To identify a potential groundwater zone, multi-parameters have been examined. Each parameter’s weights were computed based on their respective relevance in the decision-making process for determining groundwater zones. AHP’s essential properties include hierarchical formulation, which is cost-effective and time-consuming, as well as exact outcomes [8]. AHP’s adaptability for expert-based factor ordering and flexibility in reordering the rank in the case of decreased accuracy are its two main advantages. The weight of each parameter is computed by using the pairwise comparative matrix (PCM) base of judgment construction. In order to establish the weight of variables and give sub-classes a ranking in the current study, a mathematically based AHP method was employed. To remove biases in parameter ranking, a comprehensive research investigation was conducted. Together, the aforementioned elements can be used to identify a potential groundwater zone. PCM was created using a 1-to-9-point scale (Table 2). Based on each factor’s significance and impact, AHP utilized PCM to evaluate each. Weights establish the circumstance or relevance degree of variables in a hierarchy. As a consequence, the comparison matrix was used to calculate the weights of the judgment-based parameters, which were then normalized. Normalized pairwise comparative matrix (NPCM) is made up of the quotients produced by dividing the sum and each judgment value in the same columns (Table 3). The result

Estimation of Groundwater Potential Zones in Southern Dry …

189

Table 1 Description of database used for study Data used

Source of data

Data type

Scale

Well location and groundwater level

India water resources information system (https://indiawris.gov.in)

Point

–

Rainfall

India meteorological department (https://mausam.imd. gov.in/)

Grid

0.25°*0.25°

Geomorphology

Geological survey of Polygon India-Bhukosh (https://bhukosh.gsi.gov.in)

1:250,000

Geology

Geological survey of Polygon India-Bhukosh (https://bhukosh.gsi.gov.in)

1:250,000

Elevation (DEM)

SRTM (https://dwtkns.com/srt m30m/)

Grid

30 m

Satellite image

Landsat-8 (https://earthexplorer.usgs. gov/)

Grid

30 m

Lineament

Geological survey of Polygon India-Bhukosh (https://bhukosh.gsi.gov.in)

1:250,000

LULC

ESRI-ArcGIS (https://www.arcgis.com/)

Grid

10 m (resampled to 30 m)

Soil

National Bureau of Soil Survey (https://nbsslup.icar. gov.in/)

Polygon

1:500,000

was divided by total number of parameters in PCM after quotient values in each row were summed (Fig. 2). A certain degree of consistency can exist during the paired evaluation of variables in AHP. As a result, it is important to assess the pairwise evaluation’s formal consistency. Consistency level of weight of parameters is measured using consistency ratio (CR). Comparing the consistency index (CI) to the random consistency Table 2 Rating scale of Satty’s analytical hierarchical process 1/9

1/7

1/5

1/3

1

3

5

7

9

Extreme

Very Strong

Strong

Moderate

Equal

Moderate

Strong

Very Strong

Extreme

Equal

More important −−−−−−−−−−−→

Less important ←−−−−−−−−− Source Saaty [30]

190

A. B. Gireesh and M. C. Chandan

Table 3 Random consistency index based on Saaty’s scale Number of variables

5

6

7

8

9

10

12

14

15

RI

1.11

1.25

1.35

1.40

1.45

1.49

1.51

1.57

1.59

Fig. 2 Method adopted for the study

Estimation of Groundwater Potential Zones in Southern Dry …

191

index is consistency ratio CI = CI =

(λmax − n) n−1

(1)

(16.731 − 15) = 0.12 15 − 1 CR =

CR =

CI RCI

(2)

0.12 = 0.07 < 0.1 1.59

In the above equation, λmax represents principal eigenvalue of matrix and n denotes number of factors used in estimation. The consistency ratio is used to demonstrate the accuracy of weights obtained using NPCM. A judgment must be changed if the CR is larger than 0.10 even though it is considered more accurate if it is less than 0.10. The consistency ratio for this study was 0.07, which was judged suitable for further investigation. To create a groundwater potential zone map of the research region, all 15 thematic layers were integrated using weighted overlay analysis using a GIS platform. GWPI =

n 

(Xa × Y b)

(3)

i

where GWPI stands for Groundwater Potential Index, X is thematic layers’ total weight, and Y is rank of the thematic layers’ sub-class. The theme map is represented by a term (a = 1, 2, 3, … X) and thematic map classes are represented by b term (b = 1, 2, 3, … Y). The final groundwater potential zone map was divided into three categories: low, moderate, and high (Tables 4 and 5).

3 Results and Discussion Geology, Geomorphology, Slope, Aspect, Elevation, Rainfall, Lineament Density, LULC, NDVI, Profile Curvature, Plan Curvature, Soil, TWI, Stream Power Index, and Slope length are among 15 criteria that are carefully evaluated to map distinct groundwater potential areas of study area. Figure 3 shows various thematic layers considered for the study. The following section is a detailed analysis of all of these factors (Table 6).

0.33

0.33

LS

0.33

0.2

0.33

Plane curvature

Profile curvature 0.2

0.33

0.33

0.14

0.11

0.2

0.2

0.33

TWI

Rainfall

0.11

0.11

0.11

SPI

0.14

0.13

0.2

0.11

0.11

0.14

0.14

Slope

0.14

0.33

0.33

0.33

0.33

0.33

3

1

3

Soil

0.2

0.33

0.33

LULC

NDVI

0.33

3

0.33

Geomorphology 3

0.33

LD

3

Geology

1

0.33

0.2

0.11

0.11

0.11

0.11

0.11

0.2

0.2

0.33

0.33

0.33

1

0.33

0.33

0.33

3

3

3

3

3

1

0.33

0.33

3

3

3

3

3

0.2

0.33 0.2

3

3

0.33

0.33 0.33 0.33

0.14 0.33 0.11

0.33 0.33 0.14

0.2

0.33 0.33 0.14

3

5

0.33

0.33

0.33

0.33

0.33

3

1

3

3

3

5

3

5

3

0.33

0.33

0.33

0.33

0.14

0.2

1

0.33

5

0.33

0.33

3

3

3

3

3

0.33

0.33

0.33

0.33

1

5

3

7

3

3

9

7

5

5

3

3

0.33

0.33

1

3

7

3

5

3

5

9

7

9

7

9

3

3

3

3

3

9

3

7

9

9

9

5

5

5

3

0.33 1

1

3

3

3

3

7

3

3

9

5

9

7

3

0.33

3

0.2

3

0.33

0.33

5

3

5

0.2

3

1

1

0.33

0.33 0.2

0.2

0.33 0.33

3

3

3

3

3

3

9

7

9

5

LULC NDVI Plane Profile Slope Soil SPI TWI Rainfall curvature curvature

0.33 0.33 0.33

3

0.33 1

1

3

3

3

3

3

1

3

Aspect

LS

Aspect Elevation Geology Geomorphology LD

Elevation

Criteria

Table 4 Pairwise comparison matrix

192 A. B. Gireesh and M. C. Chandan

0.05

0.03

0.02

0.02

0.02

0.03

0.02

0.01

0.01

Plane curvature

Profile curvature

Slope

Soil

0.02

0.03

0.01

0.02

0.03

SPI

TWI

Rainfall

0.03

0.03

0.03

LULC

NDVI

0.05

0.05

0.05

0.03

0.03

LD

LS

0.05

0.43

0.23

Geology

Geomorphology 0.23

0.05

0.14

0.08

0.23

Aspect

0.03

0.01

0.01

0.02

0.01

0.01

0.03

0.03

0.03

0.03

0.03

0.30

0.10

0.30

0.03

0.05

0.03

0.03

0.03

0.03

0.03

0.05

0.05

0.08

0.08

0.08

0.24

0.08

0.08

0.08

LS

0.12 0.11 0.02

0.01 0.01 0.02

0.01 0.01 0.01

0.01 0.01 0.01

0.01 0.01 0.01

0.01 0.01 0.01

0.12 0.11 0.01

0.01 0.01 0.02

0.12 0.11 0.06

0.01 0.04 0.02

0.04 0.11 0.02

0.12 0.11 0.19

0.12 0.11 0.19

0.12 0.11 0.19

0.14

0.01

0.01

0.01

0.01

0.01

0.08

0.03

0.08

0.08

0.08

0.14

0.08

0.14

0.08

0.01

0.01

0.01

0.01

0.01

0.01

0.04

0.01

0.22

0.01

0.01

0.22

0.13

0.13

0.13

0.06

0.01

0.01

0.01

0.01

0.02

0.10

0.06

0.13

0.06

0.06

0.17

0.13

0.10

0.10

0.05

0.05

0.01

0.01

0.02

0.05

0.11

0.05

0.08

0.05

0.08

0.14

0.11

0.14

0.11

0.08 0.06 0.06 0.04

0.08 0.04 0.02 0.01

0.00 0.01 0.01 0.01

0.02 0.04 0.00 0.01

0.05 0.04 0.01 0.01

0.05 0.04 0.06 0.01

0.05 0.04 0.06 0.12

0.05 0.04 0.06 0.01

0.11 0.11 0.06 0.12

0.05 0.04 0.06 0.01

0.05 0.09 0.06 0.01

0.14 0.11 0.17 0.20

0.08 0.11 0.13 0.12

0.14 0.11 0.17 0.20

0.06

0.02

0.01

0.01

0.02

0.02

0.07

0.03

0.09

0.04

0.05

0.19

0.12

0.15

0.10

TWI Rainfall Weights

0.11 0.11 0.09 0.12

LULC NDVI Plane Profile Slope Soil SPI curvature curvature

0.12 0.11 0.19

Aspect Elevation Geology Geomorphology LD

Elevation

Criteria

Table 5 Normalized pairwise comparison matrix

Estimation of Groundwater Potential Zones in Southern Dry … 193

194

A. B. Gireesh and M. C. Chandan

Fig. 3 Selected thematic layers of the groundwater affecting factors a Elevation, b Aspect, c Slope, d Geology, e Geomorphology, f Lineament density, g TWI, h LULC, i NDVI, j Plan curvature, k Profile curvature, l Slope length, m Rainfall, n Stream power index, o Soil

3.1 Assessment of Groundwater Potential Zones Groundwater potential zone assessment and mapping using Geoinformatics is widely practiced in India and around the world [3]. The key to achieving reasonable results is precisely assigning weights [4]. The governing elements of groundwater potential are determined by study region and obviously, the available information. Thematic layers were generated from vector data based on these factors. The accuracy of

Estimation of Groundwater Potential Zones in Southern Dry … Table 6 Percentage of influencing factor based on Saaty’s analytical hierarchical process

195

Influencing factor

% of influence

Geomorphology

19

Elevation

15

Geology

12

Aspect

10

LULC

9

Plan curvature

7

Rainfall

6

Lineament density

5

Slope length

3

NDVI

3

Profile curvature

3

Slope

3

TWI

2

Soil

2

Stream power index

1

Table 7 Classification of groundwater potential zone Groundwater potential zone class

Area covered in km2

Area in percentage

High

3240.85

27.43

Moderate

7775.45

65.81

795.51

6.75

Low

map improves as number of layers increases [15]. The prospective map’s accuracy depends on how accurately each layer’s weight is assigned [22]. To determine potential zone in research area, 15 criteria were integrated using a GIS and AHP-based approach. In order to produce geospatial mapping of groundwater potential in the WOA, the impact of these characteristics was initially assessed using PCM and AHP. Finally, the raster map output was produced. The groundwater potential zones have been reclassified into three classes based on the weight average: high groundwater potential zone, moderate groundwater potential zone, and low groundwater potential zone. The results are shown in Table 7 and Fig. 4 illustrates the groundwater potential spatial variability.

3.2 Validation of Groundwater Potential Zone The groundwater potential map was created using a Geoinformatics and AHP-based approach using information that was accessible on several variables. However, the

196

A. B. Gireesh and M. C. Chandan

Fig. 4 Groundwater potential zone

outcomes must be verified utilizing real data that were gathered from ground. To validate generated groundwater potential maps, researchers used a number of different methods. To compare potential zone with field data, authors, [19] and [3], used depth to water level. In this investigation, which included 178 boreholes, the groundwater level was compared with groundwater potential mapped using a GIS-based approach. The groundwater data from field are divided into three categories: high level (0–5 mbgl), moderate level (5–10 mbgl), and low level (more than 20 mbgl). Table 8 shows the procedure in detail as well as comparison results. Table 8 Validation of potential zones with CGWB actual information (Sample table showing 10 reference points) Serial number

Latitude

Longitude

GWL

Zone

Evaluation

1

12.48333

76.68333

1.05

High

Positive

2

12.38889

76.68333

1.40

High

Positive

3

12.45833

76.84167

7.77

Moderate

Positive

4

12.18333

76.27500

1.45

High

Positive

5

12.81667

76.75833

2.12

High

Positive

6

12.40000

77.13333

4.80

High

Positive

7

12.45472

77.09306

4.99

High

Positive

8

12.28333

76.88333

2.54

High

Positive

9

12.33278

77.08750

7.14

High

Negative

10

12.33611

76.16111

8.13

High

Negative

Estimation of Groundwater Potential Zones in Southern Dry …

197

Based on information about groundwater levels acquired from the Central Groundwater Board, the accuracy evaluation of groundwater potential zone map was carried out. There are 178 wells with a range of levels from 0 to 19.6 mbgl. These wells were categorized into three classes: high (0–5 mbgl), moderate (5–10 mbgl), and low (0–5 mbgl) (more than 10 mbgl). Model validation of predicted groundwater potential zones with groundwater level data indicates an accuracy of 80.4%, with around 142 wells meeting the predicted groundwater potential zones. It indicates that generated groundwater potential map agrees well with ground data. The accuracy of correlation between study results and data from CGWB is acceptable. The location of groundwater-level observation point was overlain on groundwater potential zone map for validation. Following that, a scatter plot was produced between the groundwater level below ground and the related groundwater potential index. Figure 5 shows a regression plot depicting the association between GWPI and GWL. The high coefficient of determination (R2 ) obtained indicates that two variables are well correlated. The R2 value 0.8, further, implies that the suggested AHP model’s estimated GWPI and GWL values are very reliable in predicting outcome. As a result, AHP-based groundwater potential zonation was successfully used, and it may be utilized as a realistic groundwater exploring strategy. The utilization, management, and sustainability of groundwater are the three main aspects considered as essential factors to achieve SDGs. SDG 6 and 13 were included in the approach used to examine groundwater interlinkages with SDG objectives such as “Ensure availability and sustainable management of water and sanitation for everyone,” sixth Sustainable Development Goal (SDG). The emphasis on data and governance is one of the most important aspects of the SDG agenda. SDG 13 implies

Fig. 5 Scatter plot between GWPI and GWL

198

A. B. Gireesh and M. C. Chandan

to “Take immediate action to address climate change and its effects,” which is treated to be the critical goals among others. These goals were successfully addressed in this research effort. The comprehensive AHP method may be taken into consideration as a well-established resource for groundwater strategy and management that will help to conserve water in line with SDGs.

4 Conclusion In the southern dry agro-climatic region including Mysuru-Mandya, the comprehensive use of remote sensing and Geoinformatics for delineating groundwater potential zones has shown to be labor, time, and cost-effective. Based on the literature review, Geology, Geomorphology, Slope, Aspect, Elevation, Rainfall, Lineament Density, LULC, NDVI, Profile Curvature, Plan Curvature, Soil, TWI, Stream Power Index, and Slope Length are the input maps, created to estimate the groundwater potential zone for study area using weighted overlay in GIS. The 15 thematic layers were created using satellite data, geographical maps, and additional data sources. According to the results, the entire study area is divided into three groundwater potential zones: high (27.43%), moderate (65.81%), and low potential zone (6.75%). The majority of the study region lies in the moderate groundwater potential zone. The delineated groundwater potential zone map was validated using the groundwater level details about the research area given by the CGWB. The potential zone map from current study is provided to decision-makers for the optimal groundwater management. For agricultural and urban uses, the current study gives information to decision-makers for optimum groundwater management and planning. The study region is comprised of agricultural land, identified to cover larger land surface among other classes; therefore, this research will help to augment irrigation infrastructure and increase the region’s agricultural production. This research can be extended by employing sophisticated machine learning techniques to improve and achieve enhanced results. In order to support the realization of the SDGs, this study will be helpful as a baseline for planning sustainable water resource management in the southern dry agro-climatic region in the coming years.

References 1. Agarwal R, Garg PK (2016) Remote sensing and GIS based groundwater potential and recharge zones mapping using multi-criteria decision making technique. Water Resour Manage 30(1):243–260 2. Al-Shabeeb A, et al (2018) Delineating groundwater potential zones within the Azraq Basin of Central Jordan using multi-criteria GIS analysis. Groundwater Sustain Dev 7:82–90 3. Arulbalaji P, Padmalal D, Sreelash K (2019) GIS and AHP techniques based delineation of groundwater potential zones: a case study from southern Western Ghats, India. Sci Rep 9(1):1– 17

Estimation of Groundwater Potential Zones in Southern Dry …

199

4. Arunbose S et al (2021) Remote sensing, GIS and AHP techniques based investigation of groundwater potential zones in the Karumeniyar river basin, Tamil Nadu, southern India. Groundw Sustain Dev 14:100586 5. Asoka A, et al 2018) Strong linkage between precipitation intensity and monsoon season groundwater recharge in India. Geophys Res Lett 45(11):5536–5544 6. Banks D, Robins N, Robins N (2002) An introduction to groundwater in crystalline bedrock. Norges geologiske undersøkelse, Trondheim 7. Chowdhury A et al (2009) Integrated remote sensing and GIS-based approach for assessing groundwater potential in West Medinipur district, West Bengal, India. Int J Remote Sens 30(1):231–250 8. Das S (2019) Comparison among influencing factor, frequency ratio, and analytical hierarchy process techniques for groundwater potential zonation in Vaitarna basin, Maharashtra, India. Groundw Sustain Dev 8:617–629 9. Díaz-Alcaide S, Martínez-Santos P (2019) Advances in groundwater potential mapping. Hydrogeol J 27(7):2307–2324 10. Fitterman DV, Stewart MT (1986) Transient electromagnetic sounding for groundwater. Geophysics 51(4):995–1005 11. Ghorbani Nejad S, et al (2017) Delineation of groundwater potential zones using remote sensing and GIS-based data-driven models. Geocarto Int 32(2):167–187 12. Guru B, Seshan K, Bera S (2017) Frequency ratio model for groundwater potential mapping and its sustainable management in cold desert, India. J King Saud Univ-Sci 29(3):333–347 13. Hammouri N, et al (2014) Groundwater recharge zones mapping using GIS: a case study in Southern part of Jordan Valley, Jordan. Arab J Geosci 7(7):2815–2829 14. Jha MK, et al (2007) Groundwater management and development by integrated remote sensing and geographic information systems: prospects and constraints. Water Resour Manage 21(2):427–467 15. Khoshand A, Kamalan H, Rezaei H (2018) Application of analytical hierarchy process (AHP) to assess options of energy recovery from municipal solid waste: a case study in Tehran, Iran. J Mater Cycles Waste Manage 20(3):1689–1700 16. Khosravi K, et al (2018) A comparison study of DRASTIC methods with various objective methods for groundwater vulnerability assessment. Sci Total Environ 642:1032–1049 17. Krishna Kumar S et al (2012) Hydrogeochemical study of shallow carbonate aquifers, Rameswaram Island, India. Environ Monit Assess 184(7):4127–4138 18. Kumar D, Ananda Rao V, Sarma VS (2014) Hydrogeological and geophysical study for deeper groundwater resource in quartzitic hard rock ridge region from 2D resistivity data. J Earth Syst Sci 123(3):531–543 19. Kumar S, Raizada A, Biswas H, Muralidhar W, Rao KS (2016) Groundwater extraction in Karnataka and its long term implications. Indian J Econo Dev 12(4):615–628 20. Lee S, Hong S-M, Jung H-S (2018) GIS-based groundwater potential mapping using artificial neural network and support vector machine models: the case of Boryeong city in Korea. Geocarto Int 33(8):847–861 21. Lee S, Kim Y-S, Oh H-J (2012) Application of a weights-of-evidence method and GIS to regional groundwater productivity potential mapping. J Environ Manage 96(1):91–105 22. Machiwal D, Jha MK, Mal BC (2011) Assessment of groundwater potential in a semi-arid region of India using remote sensing, GIS and MCDM techniques. Water Resour Manage 25(5):1359–1386 23. Mukherjee S, et al (2017) Increase in extreme precipitation events under anthropogenic warming in India. Weather and Climate 20:1–9 24. Naghibi SA, Ahmadi K, Daneshi A (2017) Application of support vector machine, random forest, and genetic algorithm optimized random forest models in groundwater potential mapping. Water Resour Manage 31(9):2761–2775 25. Naghibi SA, et al (2015) Groundwater qanat potential mapping using frequency ratio and Shannon’s entropy models in the Moghan watershed, Iran. Earth Sci Inform 8(1):171–186

200

A. B. Gireesh and M. C. Chandan

26. Oh H-J, et al (2011) GIS mapping of regional probabilistic groundwater potential in the area of Pohang City, Korea. J Hydrol 399(3–4):158–172 27. Patra S, Mishra P, Mahapatra SC (2018) Delineation of groundwater potential zone for sustainable development: a case study from Ganga Alluvial Plain covering Hooghly district of India using remote sensing, geographic information system and analytic hierarchy process. J Clean Pro 172:2485–2502 28. Pradhan B (2009) Groundwater potential zonation for basaltic watersheds using satellite remote sensing data and GIS techniques. Central Eur J Geosci 1(1):120–129 29. Rahmati O, Pourghasemi HR, Melesse AM (2016) Application of GIS-based data driven random forest and maximum entropy models for groundwater potential mapping: a case study at Mehran Region, Iran. Catena 137:360–372 30. Saaty TL (1980) The analytic hierarchy process: planning, priority setting, resources allocation. McGraw-Hill, New York 31. Saraf AK, Choudhury PR (1998) Integrated remote sensing and GIS for groundwater exploration and identification of artificial recharge sites. Int J Remote Sens 19(10):1825–1841 32. Senanayake IP et al (2016) An approach to delineate groundwater recharge potential sites in Ambalantota, Sri Lanka using GIS techniques. Geosci Front 7(1):115–124 33. Shahid S, Nath SK, Roy J (2000) Groundwater potential modelling in a soft rock area using a GIS. Int J Remote Sens 21(9):1919–1924 34. Tahmassebipoor N, et al (2016) Spatial analysis of groundwater potential using weights-ofevidence and evidential belief function models and remote sensing. Arab J Geosci 9(1):1–18 35. Verma N, Patel RK (2021) Delineation of groundwater potential zones in lower Rihand River Basin, India using geospatial techniques and AHP. The Egypt J Remote Sens Space Sci 24(3):559–570

Evaluation and Prediction of Land Use and Land Cover Changes in the Kumaradhara Basin, Western Ghats, India N. Roopa, N. Namratha, H. Ramesh, and K. C. Manjunath

Abstract Land use land cover (LULC) is considered as the most significant and obvious indicator of changes in ecosystems. An understanding of current and potential future development opportunities is provided through analysis on the spatiotemporal shifting patterns of LULC and simulation of future scenarios. Kumaradhara river flows in the Western Ghats in southern peninsular, India. It is the major tributary of the Netravathi river, and the catchment has numerous perennial streams and is dominated by dense evergreen forests with high conservation value. In the present study, an integrated approach of remote sensing and geospatial techniques is used to assess LULC changes for the period of 2010–2020, and prediction of future LULC change was carried out by ANN model using MOLUSCE plugin of QGIS for the year 2025. The results have shown that the build-up land has increased considerably, and forest has decreased which is evident from the increase in cultivated land. The predicted LULC showed an increase in built-up land and a significant transformation of barren land. The results of this study indicate significant changes in the LULC pattern. Keywords Land use/Land cover · Predicted LULC maps · MOLUSCE (ANN multilayer perception model) · Kumaradhara basin

N. Roopa · H. Ramesh Department of Water Resources and Ocean Engineering, National Institute of Technology Karnataka, Surathkal, India N. Namratha (B) Department of Hydraulics Engineering, The National Institute of Engineering, Mysuru, India e-mail: [email protected] K. C. Manjunath Department of Civil Engineering, The National Institute of Engineering, Mysuru, India © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_15

201

202

N. Roopa et al.

1 Introduction The landscape structure or land cover of a river basin’s watershed determines its hydrological yield. Anthropogenic and natural forces both contribute to the current changes in the earth’s land surface [1]. These factors have an impact LULC change directly or indirectly. Anthropogenic activities include deforestation, urbanization, overgrazing, and conversion to agricultural land. These are the primary proximal reasons for LULC change [2]. Earthquakes, landslides, droughts, and floods are the natural factors that affect the land use and land cover [3]. LULC changes affect an area’s climate, which in turn impacts on natural resources such as water, wetlands, and biodiversity [4]. It is also known to have an impact on the hydrology of any catchment area. The LULC patterns are greatly influenced by climatic factors such as rainfall, temperature, and these factors need long-term research to fully comprehend the trend. LULC change detection is a complex function and requires different methods of analysis based on spatial and social data in order to understand the drivers and impacts of change with time. The LANDSAT imagery archive runs from 1973 to the present (LANDSAT 8) and contains the most recent satellite data that are reasonable with prior collective years [5, 6]. Remote Sensing and Geographic Information Systems (GIS) when compared to ground-based surveys are the most cost-effective techniques for mapping and assessing LULC changes [7, 8]. Modules of land-use change evaluation (MOLUSCE), a Quantum GIS (QGIS) plugin is used by most of the researchers to assess prospective LULC changes using CA model and also contains a probability of a transition matrix [9]. Western Ghats (WG) of India is one of the 34 global biodiversity hotspots and habitat loss is a wreaking havoc for the past few decades [10]. Along the west side of the Western Ghats, the Kumaradhara river is a significant feature of rugged forested terrain. It is the major tributary of the Netravathi river and the catchment has numerous perennial streams and is dominated by dense evergreen forests with high conservative value [11]. The availability of relevant historical satellite datasets makes quantifying LULC change very simple, however, identifying the causes for such change is difficult. The climate, magnitude, time, ecological, social, political institutions, economies, and international protocols all influence the drivers [12]. Studying the land change pattern and its mechanism will help one to further analyze the effects of changes in the hydrology of a catchment, which affects various hydrological cycle metrics. The aim of the present study is to adopt geospatial and remote sensing techniques to evaluate and track LULC shifts for three different years in the Kumaradhara river basin. The present paper attempts to evaluate LULC changes between various classes from 2010 to 2020, forecast LULC change, and prepare a prediction map for 2025. The study also highlights the importance of evaluating LULC change and its applications for further studies.

Evaluation and Prediction of Land Use and Land Cover Changes …

203

2 Materials and Methodology 2.1 Study Area Kumaradhara river is the tributary of the Netravathi river, which has its origin in Western Ghats of Karnataka, flowing to a length of 126 km and joins the Arabian Sea on the west. Kumaradhara river basin extends from 12°29' 4'' N to 12°58' 33'' N latitude and 75°95' 8'' E to 75°47' 48'' E longitude with the catchment area of 1776 km2 , originates in Central Western Ghats at an elevation of 1480AMSL [10]. Kumaradhara basin is rich in biodiversity and the freshwater ecosystems are extremely diverse, distinctive, and vital to livelihoods and economies. It spreads across three districts, Hassan, Kodagu, and Dakshina Kannada. The catchment is shown in Fig. 1. Kumaradhara river has tributaries, namely, Hongadahalla and Kadumanehalla rivers, and sub-catchments covering these parts of the Mookanamane and Marenahalli rivers are chosen for the present study. Mookanamane catchment is flanked by Bisle and Kaginele forests on either side and covers an area of 42.24 km2 , which extends from 12°45' 0'' N to 12°49' 30'' N latitude and 75°43' 30'' E to 75°46' 30'' E longitude with an elevation of 1179 m AMSL, the annual average rainfall in this area is 2978 mm, and annual average maximum and minimum temperatures are 28 °C and 19 °C, respectively. Marenahalli catchment covers an area of 64.19 km2 extending from 12°52' 30'' N to 12°57' 30'' N latitude and 75°40' 0'' E to 75°46' 30'' E longitude with an elevation of 1293 m AMSL, the annual average rainfall, maximum, and

Fig. 1 Study area

204

N. Roopa et al.

minimum temperatures are 3439.4 mm, 29 °C, and 26 °C. The study area mainly covers forests and landscapes associated with Kumaradhara river.

2.2 Data Used in the Study Remotely sensed LANDSAT data were used to discriminate between the various LULC cover categories. The fluctuations in the Kumaradhara river catchment were monitored using the pictures acquired from the USGS Earth Explorer website. The ETM+ sensor of LANDSAT 7 can produce images for 2010 and 2015, and LANDSAT 8 Operational Land Imager (OLI) and TIRS CI-Level 1 sensor can produce images for 2020. For the months of February and March in 2010, 2015, and 2020, all these images were downloaded. Due to the region’s harsh weather conditions and significant seasonal variations in LULC, especially vegetation, these months were chosen for data collection. The following characteristic data were picked for the study, digital elevation model (DEM), slope maps, aspect maps, and distance from road maps. The road map was downloaded from the website of open street maps. Table 1(a) and 1(b) provides specifics of the original source of datasets. Table 1 a Source of datasets. b LANDSAT data used for the study Criteria

Description

Source

Digital elevation model

DEM Gradient Aspect

SRTM with 30 m spatial resolution

National remote sensing centre/ Bhuvan/DEM/ elevation

Road map

Distance from road map Road map of Kumaradhara river basin

Data (a): Source of datasets

Year

Satellite ID

Sensor ID

Path

Row

Open street map

Acquisition date

(b): LANDSAT data used for the study 2010

LANDSAT 7

ETM+

145

051

18/02/2010

2015

LANDSAT 7

ETM+

145

051

27/03/2015

2020

LANDSAT 8

OLI and TIRS

145

051

09/03/2020

Evaluation and Prediction of Land Use and Land Cover Changes …

205

2.3 Methodology 2.3.1

Preparation of LULC Maps

The LULC classification method depends on the nature of the study area, spatial, and temporal resolution of the data [13]. The LANDSAT images downloaded from USGS website are georeferenced to Universal Traverse Mercator (UTM) projection system, 43° N WGS 84, which was used throughout the study. The satellite images are imported to QGIS and False-color composite (FCC) was created using the bands 4, 3, and 2 for LANDSAT 7 and bands 5, 4, and 3 for LANDSAT 8. Radiometric corrections are applied to reduce the inconsistencies of the image and to enhance the quality of the satellite images.

2.3.2

Classification and Accuracy Assessment

The LULC maps are classified into five different classes, water bodies, forest, cultivated land, built-up land, and barren land, using the LANDSAT level 1 classification in the present study. A maximum likelihood algorithm is used to determine the number of land cover types and the training pixels for each of the desired classes using land cover information. The description of the land use classes and land cover statistics is given in Table 2. The adopted equations for the calculation of user’s accuracy, producer’s accuracy, overall accuracy, and kappa coefficient are as follows, ( Users’ accuracy =

xi xj

) ∗ 100

where, xi = Number of correctly classified samples in each category xj = Total number of reference samples in that category (Row table) ( Producers’ accuracy =

xi xk

) ∗ 100

where, xk = Column total of the reference samples ( Overall accuracy =

xd xT

) ∗ 100

where, xd = Diagonal total of correctly classified samples xT = Total number of reference points

206

N. Roopa et al.

( Kappa coefficient =

(xd ∗ x T ) − B (x T ∗ x T ) − B

)

where, B = Sum of products of row total and column total for each LULC type in confusion matrix. In the present study, a confusion matrix was produced to calculate the overall accuracy and kappa coefficient for each land use land cover category.

2.3.3

Prediction Map of 2025

The cellular automaton model describes the LULC modifications using the probabilities obtained from the transitions through ANN learning process [14]. MOLUSCE is computing land use change analysis effectively [15], and future LULC maps will be created using the plugin QGIS 2.18.24. The correlations among the spatial variable parameters are calculated using the Pearson’s correlation, and then the change in LULC between the initial year (2010) and the final year (2020) is calculated. The methodology is shown in Fig. 2.

3 Results and Discussion 3.1 Comparison of Landcover Datasets The study area was classified according to the requirements of the research work and its applications. Marenahalli catchment spreads over 64.19 km2 . The classification maps for the years 2010, 2015, and 2020 are shown in Fig. 3. Table 3 explains the percentage of area covered by each class in the Marenahalli catchment. According to 2010 results, the maximum area of the catchment, i.e., 93.79% (60.18 km2 ) of the area is covered by forest. In the year 2015, it covers 92.94% (59.67 km2 ) and in 2020 92.61% (59.43 km2 ). Mookanamane catchment spreads over 42.24 km2 in which 89.25% (36.80 km2 ) of the total area is covered by forest in the year 2010, and in 2015 and 2020, the area covered is 87.64% (35.39 km2 ), and 87.22% (35.22km2 ). The classification maps of Mookanamane are shown in Fig. 4, and Table 4 explains the percentage of area covered by each class. The forest covers in these two catchments are the lands with tree canopy, evergreen, semi-evergreen vegetation types producing other forest products. Built-up land is the area developed due to non-agricultural uses and barren lands are the degraded lands devoid of vegetation.

Evaluation and Prediction of Land Use and Land Cover Changes …

Fig. 2 Flowchart of methodology

Fig. 3 LULC maps of Marenahalli, 2010 (a), 2015 (b), 2020 (c)

207

208

N. Roopa et al.

Table 2 Land use land cover categorization scheme for the study area Serial number

Class

Description

1

Water bodies

Surface water that is retained as ponds, rivers, flowing as streams and other bodies of water

2

Forest

Land covering dense or evergreen forests, deciduous forest, tall grass, and scrublands

3

Cultivated land

A cycle of a cultivated area, involving a brief period of farming, harvest and then a return to bare soil

4

Built-up land

The land occupied with buildings and other man—made constructions and also settlements such as roads, industries etc.,

5

Barren land

Land having lack of water, soil management, and sparsely vegetated terrain and Bare rock exhibits evidence of erosion, and ground deformation

Fig. 4 LULC maps of Mookanamane, a 2010, b 2015, c 2020 Table 3 LULC classification of Marenahalli LULC class

2010 km2

Water bodies Forest Cultivated land

2015 %

km2

2020 %

0.019

0.030

0.027

0.042

60.185

93.779

59.672

92.940

0.006

0.009

0.012

0.018

km2 0.015 59.43 0.019

% 0.030 92.57 0.030

Built-up land

0.931

1.451

1.208

1.780

1.469

2.290

Barren land

3.035

4.730

3.257

5.207

3.235

5.090

Evaluation and Prediction of Land Use and Land Cover Changes …

209

Table 4 LULC classification of Mookanamane LULC class

2010 km2

Water bodies Forest Cultivated land

2015 %

km2

2020 %

km2

%

0.015

0.036

0.012

0.020

0.008

0.010

36.808

89.251

35.396

87.640

35.277

87.220

0.034

0.083

0.053

0.129

0.039

0.100

Built-up land

0.086

0.210

0.136

0.322

0.197

0.480

Barren land

4.296

10.418

5.246

11.870

5.627

12.190

3.2 Accuracy Assessment and Change Analysis of Classified Images Accuracy assessment and change detection of land use land cover are very important for understanding the interaction and relationship between humans and nature [16]. The accuracy of LULC is calculated using various measures such as overall accuracy, user’s accuracy, producer’s accuracy, and Cohen’s kappa coefficient [17]. In this study, a confusion matrix was produced to calculate the overall accuracy and kappa coefficient for each land use and land cover category. The value of overall accuracy and kappa coefficient is given in Table 5. According to Landis and Koch [18], Cohen’s Kappa shows substantial and perfect agreement if the values fall between 0.6 and 1. In the present study, the kappa coefficients of both the catchments lie in the above range and hence it is considered as substantial agreement.

3.3 LULC Change Detection The LULC classes and its changes of Marenahalli catchment in the year 2010– 2015 and 2015–2020 are shown in Table 6. Water bodies and barren land exhibited both negative and positive changes whereas the cultivated land and built-up land showed positive changes, but it is observed that forest area is decreasing with negative changes. Minimal changes were observed in all other classes and built-up land is having the most substantial changes. The forest covers the maximum area of the catchment and it showed a decreasing trend. The decrease in forest is attributed to conversion of forest lands into developed area being used for various types of developmental works. The area under built-up land increased due to gradual increase in demand for shelter by inhabitants in and around Marenahalli. The area of barren land is rising due to farmer’s disregard of agricultural operations and clearing of land, which suggests a clear plan for new structures and communities. The changes in LULC classes for the years 2010–2015 and 2015–2020 are presented in Table 7 for Mookanamane catchment built-up land, barren land, and cultivated land shows an increasing trend. Similar to Marenahalli, the forest area

210

N. Roopa et al.

Table 5 Accuracy assessment Catchment

Year

Overall accuracy

Kappa coefficient

Marenahalli

2010

91.00

0.88

2015

91.00

0.90

2020

89.04

0.86

2010

91.02

0.88

2015

88.05

0.85

2020

89.00

0.86

Mookanamane

Table 6 Change analysis in Marenahalli Year

2010–2015

2015–2020

Area in km2 2010 Water bodies Forest

Changes in %

2015

Area in km2 2015

0.019

0.027

+0.008

0.027

60.185

59.672

−0.513

59.672

Changes in %

2020 0.015 59.43

−0.012 −0.242

Cultivated land

0.006

0.012

+0.006

0.012

0.019

+0.007

Built-up land

0.931

1.208

+0.277

1.208

1.469

+0.261

Barren land

3.035

3.257

+0.222

3.257

3.235

−0.022

is decreasing. Built-up land shows an increasing trend due to gradual increase in communities of people. It has been recognized as a tourist spot due to which there is an increase in buildings, roads, and hotels. Forest areas are reducing due to human intrusion at forest areas and other settlement activities.

3.4 Forecasting Future LULC Changes Forecasting LULC changes is important for recognizing and emphasizing probable future adjustments and modifications to the landscape [19]. The prediction is supported by the previous LULC patterns seen in the study area. In the present study, a prediction of LULC change for the year 2025 has been attempted. The variables used for the prediction are LULC maps, slope maps, aspect maps, and the distance from the road maps. The trends of the historical changes have been used for the prediction of future LULC maps. Tables 8 and 9 show the predicted land area for the year 2025 (Fig. 5).

Evaluation and Prediction of Land Use and Land Cover Changes …

211

Table 7 Change analysis in Mookanamane 2010–2015

Year

2015–2020

Area in km2 2010

2015

Area in km2 2015

Changes in %

2020

0.015

0.012

−0.003

0.012

0.008

−0.004

36.808

35.396

−1.412

35.396

35.277

−0.119

Water bodies Forest

Changes in %

Cultivated land

0.034

0.053

+0.019

0.053

0.039

−0.014

Built-up land

0.086

0.136

+0.050

0.136

0.197

+0.061

Barren land

4.296

5.246

+0.950

5.246

5.627

+0.381

Table 8 Distribution of LULC classes in 2025 at Marenahalli LULC class 2010 km2 Water bodies Forest

0.019 60.18

2015 % 0.030 93.77

km2 0.027 59.67

2020 % 0.042 92.94

km2

2025 %

0.015 59.43

0.030 92.57

km2 0.014 59.30

% 0.022 92.40

Cultivated land

0.006

0.009

0.012

0.018

0.019

0.030

0.021

0.033

Built-up land

0.931

1.451

1.208

1.780

1.469

2.290

1.571

2.448

Barren land

3.035

4.730

3.257

5.207

3.235

5.090

3.305

5.150

Table 9 Distribution of LULC classes in 2025 at Mookanamane LULC class 2010 km2 Water bodies Forest

0.015 36.80

2015 % 0.036 89.25

km2 0.012 35.39

2020 % 0.020 87.64

km2 0.008 35.27

2025 % 0.010 87.22

km2 0.015 35.14

% 0.036 85.22

Cultivated land

0.034

0.083

0.053

0.129

0.039

0.100

0.0776

0.19

Built-up land

0.086

0.210

0.136

0.322

0.197

0.480

0.2569

0.62

Barren land

4.296

10.418

5.246

11.870

5.627

12.190

5.7609

13.97

212

N. Roopa et al.

(1)

(2)

Fig. 5 Prediction map of (1) Mookanamane and (2) Marenahalli

4 Conclusion Kumaradhara catchment is rich in biodiversity and is a major source of varieties of species, flora, and fauna. The present study investigates LULC changes for sustainable development through an object-oriented image classification approach. This study has revealed that there are changes in LULC categories over the year (2010–2020), which is presented in Tables 6 and 7. The results of the change detection in Marenahalli catchment showed that the largest change was recorded in the forest land cover class −0.853% for a span of 10 years from 2010 to 2020. The barren land and built-up land were the next dynamic land use categories. The area under barren land increased to 0.22% and again decreased 0.022% from 2015 to 2020. Built-up land increased 0.48% from the year 2010 to 2020. Cultivated land also increased considerably but at a very lower rate. The predictions of future LULC change also revealed that the trends in decreasing forests are going to continue in future, and there will be minimal increase in builtup and cultivated lands. A substantial amount of preservation and conservation is required for the forest area in and around Marenahalli catchment, which is a crucial area. Marenahalli LULC classified images show an overall accuracy of 91% in 2010, 91% in 2015, and 89% in 2020. The change detection results of the Mookanamane catchment revealed that the forest land cover class saw the higher level of change −1.58% from the year 2010– 2020. The other land use classes that increased were barren land at the rate of 0.95% and built-up land at the rate of 0.06%. Although at much slower rate, the cultivated land expanded from the year 2010 to 2015 but gradually it showed a decreasing trend −0.014% from the year 2015 to 2020. This catchment showed a considerable drop

Evaluation and Prediction of Land Use and Land Cover Changes …

213

in land use classes such as forest and the water bodies and an increase in barren and built-up land. The projections of future LULC change also showed that there will be a significant increase in built-up and cultivated regions, as a continuation of trends towards decreasing woods. The forest region in and around the Mookanamane watershed decreases significantly. Also, Mookanamane LULC classified images show an overall accuracy of 91% in 2010, 88% in 2015, and 89% in 2020. Forecasted LULC classes are shown in Tables 8 and 9. Changes in land use and cover become a key part of modern resource management and environmental change monitoring systems. Natural resources such as water, soil, nutrients, plants, and animals are significantly impacted by land use and land management strategies. Information on land usage may be utilized to provide solutions for problems with managing natural resources like salinity and water quality. The study of historical LULC change in Marenahalli and Mookanamane catchment will be helpful to evaluate the relation between LULC change, and the results of this study can be used for further watershed hydrology studies through advanced modeling tools. In conclusion, the present study highlights the scarcity of research conducted in the aforementioned study area concerning land use land cover (LULC) analysis. By assessing and predicting LULC changes, this paper contributes valuable insights for future investigations, aimed at establishing correlations with hydrological parameters, thereby paving the way for enhanced resource management and environmental change monitoring in the region.

References 1. Aspinall RJ (2008) Land use change, science, policy and management. CRC Press, Taylor and Francis group 2. Lambin EF (2008) Land- use and land-cover change local processes and global impacts. Springer, Berlin Hyderberg New York 3. Gupta S, Roy M (2012) Land use/ land cover classification of an urban area- a case study of Burdwan municipality, India. Int J Geomatics Geosci 2:1024–1036 4. Pachauri R (2007) Climate change, 2007, synthesis report. IPCC 5. Hakim BS, Arif S (2019) Spatial dynamic prediction of land use/land cover change (Case study: Tamalanrea sub district, Makassar city. IOP conference Earth Environmental science. Institute of Physics 6. Rawat JS, Biswas V (2013) Changes in land use /cover using geospatial techniques: a case study of Ramnagar town area, district Nainital, Uttarakhand, India. Egypt J Remote Sens Space Sci 16(1):111–117 7. Subhashini S, Thirumaran K (2016) A comparative analysis of land surface retrieval methods using Landsat 7 and 8 data to study urban heat island effect in Madurai. Int J Earth Sci Eng 09(04):1397–1404 8. Srivastava R (2020) Changes in vegetation cover using GIS and remote sensing: a case study of south campus BHU, Mirzapur, India. J Sci Res 64:135–141 9. Saputra MH, Lee H (2019) Prediction of land use and land cover changes for North Sumatra, Indonesia, using an artificial-neural-network-based cellular automaton. Sustainability 11 10. Ramachandra TV (2013) Kumaradhara River Basin, Karnataka Western Ghats: need for conservation and sustainable use. Environmental Information System [ENVIS]

214

N. Roopa et al.

11. Rakesh Kumar Sinha TE (2018) Effects of historical and projected land use/cover change on runoff and sediment yield in the Netravati river basin, Western Ghats, India. Environ Earth Sci 77(3):1–19 12. Mangueira JR, Vieira LT, Azevedo TN (2021) Plant diversity conservation in highly deforested landscapes of the Brazilian Atlantic Forest. Perspect Ecol Conserv 19(1):69–80 13. Kaufmann R, Seto KC (2001) Change detection, accuracy, and bias in a sequential analysis of Landsat imagery in the Pearl River Delta, China: econometric techniques. Agric, Ecosyst Environ 85:95–105 14. NextGIS (2017) MOLUSCE—quick and convenient analysis of land cover changes. https:// nextgis.com/blog/molusce/ 15. Sathya BA, Shashi M (2020) Future land use land coverscenario simulation using open source GIS for the city of Warangal, Telangana, India. Appl Geomatics 12:281–290 16. Chughtai AH, Abbasi H (2021) A review on change detection method and accuracy assessment for land use land cover. Remote Sens Appl: Soc Environ 22:100482 17. Congalton RG (2019) Assessing the accuracy of remotely sensed data: principles and practices, 3rd Edition. CRC Press, India 18. Landis JR, Koch GG (1977) An application of Hierarchical kappa-type statistics in the assessment of majority agreement among multiple observers. Biometrics 33(2):363–374 19. Marwa Waseem A, Halmy PE (2015) Land use/land cover change detection and prediction in the north-western coastal desert of Egypt using Markov-CA. Appl Geogr 63:101–112 20. Bangalore D (2013) S Land use and land cover mapping by using remote sensing and GIS techniques—a case study of Kasaba Hobli, Hoskote Taluk, Bangalore Rural District, Karnataka, India. Int J 2(1):1–6

Evaluation of Surface Soil Moisture Using Remote Sensing and Field Studies T. N. Santhosh Kumar and Abhishek A. Pathak

Abstract Soil moisture (SM) is an important quantity to examine in terms of agriculture, meteorology, and hydrology to understand the evaporation cycle and drought mechanisms. This study aims to estimate surface soil moisture in arid areas using Sentinel-1A SAR data. In order to collect soil samples from sampling grids that are synchronized with Sentinel-1A passes, study area is divided into 80 grids, each measuring 10 m by 10 m. Six SAR images were collected from Copernicus Open Access Hub website. The vegetation index (NDVI) was calculated using a Sentinel2A image. The SNAP software was used to process the SAR images, and R studio was used to extract NDVI values and backscattered energy of each sample grid. In this study, an empirical equation was developed to model surface soil moisture using the dielectric constant and backscattering coefficients. The performance of the model was assessed using statistical indicators such as the coefficient of correlation, Nash–Sutcliffe efficiency, and root mean square error, which yielded results of 0.85, 1.46, and 0.75, respectively. Keywords SAR data · Soil mapping · Semi-empirical model · Bare field

1 Introduction Surface soil moisture (SSM) is a crucial element that influences hydrological cycle, ecology, and energy fluxes between ground and atmosphere [16]. SSM data are critical in many industries, including agriculture, meteorology, and hydrology [19]. Soil moisture is one of the least monitored of all the hydrologic variables. It is greatly influenced by unpredictable and intermittent precipitation, varying evapotranspiration rates, heterogeneous soils, land cover, and topography and is extremely changeable in both space and time [15]. Monitoring soil moisture, thus, becomes quite challenging. T. N. S. Kumar · A. A. Pathak (B) Department of Civil Engineering, The National Institute of Engineering, Mysuru, India e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_16

215

216

T. N. S. Kumar and A. A. Pathak

In order to study global behavior and earth changes, remote sensing is a valuable instrument. Moreover, for determining distribution and amount of soil moisture at various scales, remote sensing provides an affordable, efficient, and time-consuming alternative to in situ measurements [10, 17]. Different remote sensing techniques have demonstrated their capacity to retrieve soil moisture from the earth’s surface, including optical, thermal, and microwave. However, microwave remote sensing has a number of advantages, including high penetration and all-weather capabilities [12, 22]. Numerous studies have been published to assess soil moisture content using X-band [1, 24], L-band [2, 7, 16, 17] and C-band [3, 11, 14, 16, 20, 21, 23], SAR images. From the available literature, it is known that soil moisture can be obtained with reasonable accuracy using C-band SAR over barren and grasslands. Several approaches for estimating soil moisture across bare soil surfaces using SAR measurements have been developed, such as physical model, advanced integral equation model, integral equation model for multiple scattering, and empirical and semi-empirical models [13], and various approaches assumed that the connection between surface soil moisture and SAR backscattering energy is linear [5, 6, 18]. The main objective of the study is to develop an empirical model for surface soil moisture using the dielectric constant and backscattering coefficients. This study uses Sentinel-1A data to estimate and map the surface soil moisture in the study region on a plot scale.

2 Study Area An agricultural plot in Mysuru district was chosen as the study area. To identify whether the study area is grassland or barren, a reconnaissance study was first conducted. The study area consists of two plots, which consist of two different soil types, i.e., sandy loam and silty clayey. Location of study area (Fig. 1a) along with Google Earth image of study area and the sample grids are shown in Fig. 1b. In these sites, a drip irrigation is practiced to irrigate the agricultural lands. Therefore, to regulate irrigation water in this area, investigations of soil moisture are essential.

3 Data 3.1 Sentinel-1A The European Space Agency (ESA) launched Sentinel-1A satellite on 3 April 2014. The SAR payloads utilize a C-band frequency of 5.405 GHz and have various operational modes, such as Interferometric wide swath, stripe map, extra wide swath, and Wave. These payloads have a 12-day repetitive cycle and capture data in dualpolarization mode (VV + VH). To obtain Sentinel-1A data for this study, the

Evaluation of Surface Soil Moisture Using Remote Sensing and Field …

217

Fig. 1 a Study area map. b Google Earth image of study area along with sampling grids

Copernicus Open Access Hub is utilized (https://scihub.copernicus.eu/dhus/#/home). Table 1 provides an overview of Sentinel-1A data features, and Table 2 lists the dates on which data were collected.

218

T. N. S. Kumar and A. A. Pathak

Table 1 Sentinel-1A features Frequency band

C-band

Resolution

(10*10 m)

Frequency

5.405 Ghz

Satellite revisit period

12 days

Polarization

VV and VH (dual-pol)

Period

2014–present

Orbit direction

Descending

Swath (km)

100–150

Table 2 Data acquisition dates Number of pass

1

2

3

4

5

6

Acquisition dates 21/3/2022 02/04/2022 26/04/2022 8/05/2022 20/05/2022 01/06/2022

3.2 Sentinel-2A The Copernicus Open Access Hub is used to acquire Sentinel-2A L1C data. Sentinel2A is an optical data collecting mission with a 10-day return time that intends to document changes in land surface conditions and assist surface modifications. It has a resolution of 10*10 m.

3.3 Field Data Two sets of soil samples were gathered from the field. The first set of samples, taken from a depth of 0–5 cm, was intended to determine the gravimetric moisture content. The second set of samples, collected using a core cutter, was taken to interpret the gravimetric moisture content as volumetric moisture content and to assess the soil texture in accordance with IS 2720–4, 1985 [9]. Soil samples were collected during each satellite pass over study region from 21 March 2022 to 1 June 2022, with the temporal resolution of 12 days (Fig. 2). The surface roughness of the soil is measured using a roller chain.

4 Method Figure 3 illustrates the overall methodology. The following sections describe the image pre-processing, field data analysis, and the semi-empirical model used in this study.

Evaluation of Surface Soil Moisture Using Remote Sensing and Field …

Fig. 2 Collecting soil samples on-site and drying them in an oven

Fig. 3 Flowchart of methodology

219

220

T. N. S. Kumar and A. A. Pathak

Fig. 4 NDVI map

4.1 Image Processing Sentinel Application Platform (SNAP) software was used to pre-process Sentinel-1A images. It contains standard procedures such as importing data, applying orbit file, radiometric calibration, multi-looking, speckle filter, geometric terrain correction, and data conversion (linear to dB). Sentinel-2A L1C data were used to develop NDVI map. In this study, single image that was downloaded on the date 22 March 2022 was used since plant growth is a variable that changes slowly. For each sample grid, the normalized difference vegetation index (NDVI) data are generated using NDVI processor tool in SNAP. NDVI map of study site is shown in Fig. 4. Using R software, backscattered value and resultant NDVI values were extracted from each sample grid.

4.2 Data Collection from Field Soil samples are taken throughout all sampling grids at 6:30 a.m. ± 1/2 hour, synchronized with satellite passes. During a single satellite visit, 80 samples were collected from the study site. To calculate the gravimetric moisture content, topsoil (0–5 cm) is collected and labeled in zip-lock packets. As per IS 2720–4, 1985 [9], the process of collecting soil samples involves using a core cutter and subsequently subjecting them to laboratory analysis for texture and bulk density. The samples are transported

Evaluation of Surface Soil Moisture Using Remote Sensing and Field …

221

to the laboratory, weighed to record their initial weight, and then dried in an oven at a temperature of 105 °C to determine the gravimetric soil moisture. After the samples have been oven-dried, their dry weights are measured, and an equation is used to calculate soil moisture is: ) ( wω − wd × 100 (1) Soil moisture = wd where, wω = wet weight of soil, wd = dry weight of soil. A roller chain approach was used to measure the Surface Soil Roughness (SSR) at the location. The statistical metric RMS (vertical variation) was used to explain SSR [16], which was obtained using Eq. (2). /( Σ RMS =

(h i − h)2 n−1

) (2)

where, h = mean height of the soil surface, hi = soil surface height, and n = number of samples

5 Analysis 5.1 Field Data Analysis The barren area was chosen to study how soil moisture affects backscattered energy in the absence of vegetation. Figure 5 shows the spatiotemporal change of soil moisture within the plot. The box plot shows that soil moisture on 20 May 2022 has increased due to the high rainfall during that period. The soil moisture on 2 April 2022, 26 April 2022, and 1 June 2022 has decreased due to less or no rainfall at that period.

5.2 Soil Moisture Retrieval Model A semi-empirical model was developed using multi-linear regression analysis. The main factors affecting the backscattered value are used to predict the surface soil moisture. Since it is a barren field, NDVI indicates no correlation, hence it is ignored. Many studies have proposed empirical models to estimate soil moisture, and these models demonstrated a basic relationship between surface soil moisture and backscattering energy [7, 14, 16]. The backscattering energy and dielectric constant were considered to model the soil moisture, and these are presented in Eqs. (3) and (4).

222

T. N. S. Kumar and A. A. Pathak

Fig. 5 Variation of soil moisture within the plot over the study period

( o ) o Mv = f σvv , σvh ,ε

(3)

o o Mv = a + b ∗ σvh + c ∗ σvv + d∗ ε

(4)

where, σo vh and σo vv are the backscattering coefficient, ε = dielectric constant.

6 Results and Discussion 6.1 Surface Soil Roughness and Backscattering Energy As described in the prior section, the surface roughness of the study area was established. Figures 6 and 7 examine the connection between surface roughness and the backscattered coefficients. The VH and VV polarizations showed minimal correlation with SSR, with R2 values of 0.01 and 0.02, respectively, indicating that backscattered energy is only marginally affected by SSR. Hence, it is neglected.

Evaluation of Surface Soil Moisture Using Remote Sensing and Field … Fig. 6 Relationship between R.M.S and VV polarizations

223

r.m.s

R² = 0.01

-15

-13

-11 -9 Sigma Naught VV

-7

0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

-5

R² = 0.02

Fig. 7 Relationship between R.M.S and VH polarizations

r.m.s

0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

-25

-20 -15 Sigma Naught VH

-10

6.2 Vegetation and Backscattering Energy Sentinel-2A data were used to generate the NDVI values of each sampling grid. NDVI and backscattered coefficients of VV and VH polarizations are analyzed in Figs. 8 and 9. Correlation study shows that there is a weaker association between NDVI and backscattered energy, with R2 = 0.01 for σ°vv and R2 = 0.15 for σ°vh . NDVI values are within the range, i.e., NDVI < 0.3. As a result, it is not used to model soil moisture.

6.3 Dielectric Constant and Backscattering Energy Field soil moisture and texture data were used to compute the dielectric constants of each sample. Dry soil has a dielectric constant of roughly 3, while water has a

224

T. N. S. Kumar and A. A. Pathak

R² = 0.01

Fig. 8 Relationship between NDVI and VV polarizations

0.3 0.25

NDVI

0.2 0.15 0.1 0.05 0

-11

-10

-9 -8 Sigma naught VV

-7

R² = 0.15

Fig. 9 Relationship between NDVI and VH polarizations

0.3 0.25 NDVI

0.2 0.15 0.1 0.05 0

-21

-20

-19 -18 -17 -16 Sigma naught VH

-15

-14

dielectric constant of 80 [4]. As a result, the water-to-soil ratio fluctuates between these ranges. According to [8], for a certain microwave frequency, dielectric constant may be determined using Eq. (5). ( ) ε = (x0 + x1 s + x2 c) + y0 + y1 s + y2 c Mv + (z0 + z1 s + z2 c) M2v

(5)

where, ε is dielectric constant, s denotes sand percentage, c stands for clay percentage, Mv represents volumetric moisture content and xi, yi, zi are frequency-dependent coefficients adopted from [8]. Backscattered energy is linearly dependent on dielectric constant, as shown in Figs. 10 and 11. R2 for VH polarization is 0.27 and 0.40 for VV polarization.

Evaluation of Surface Soil Moisture Using Remote Sensing and Field …

R² = 0.40 25

Dielectric constant

Fig. 10 Relationship between dielectric constant and VV polarization

225

20 15 10 5 0

-15

-11 -9 Sigma naught VV

-7

-5

R² = 0.27 Dielectric constant

Fig. 11 Relationship between dielectric constant and VH polarization

-13

25 20 15 10 5 0

-25

-20 -15 Sigma naught VH

-10

6.4 Soil Moisture and Backscattering Energy For data obtained at 12-day intervals from 21 March 2022 to 1 June 2022. The correlation between volumetric moisture content and backscattered energy is portrayed in Figs. 12 and 13. With an R2 of 0.45 for VV polarization and 0.34 for VH polarization, backscattered energy is strongly dependent on soil moisture. The findings indicate that the backscattering coefficient of VV polarization has the strongest correlation with soil moisture as observed in the field. This is primarily due to the fact that co-polarization has superior penetration capabilities compared to cross-polarization (Rawat et al. 2017).

226

T. N. S. Kumar and A. A. Pathak

R² = 0.45 60

Volumetric moisture

Fig. 12 Relationship between volumetric moisture content and VV polarization

50 40 30 20 10 0

-15

-11 -9 Sigma naught VV

-7

-5

R² = 0.34 60 50 40 30 20 10 0

Volumetric moisture

Fig. 13 Relationship between volumetric moisture content and VH polarization

-13

-24

-22

-20 -18 -16 -14 Sigma naught VH

-12

-10

6.5 Model Development Major factors that influence the sensitivity of backscattered energy are used to model the surface soil moisture. Multilinear regression analysis is used to develop a model, with volumetric moisture content (Mv) as the dependent variable and backscattered energy, dielectric constant as the independent variables. The proposed model has a 95% confidence level with an adjusted R2 of 0.85 and RMSE of 2.11. Equation 6 describes the empirical model that was developed. The individual association between Mv and other dependent variables is moderate; which manifests the significance of considering the multivariable in modeling SSM. Similar results can be observed in Gururaj et al. [7]. Mv = −1.8417 − 0.71564 ∗ σ o vh + 0.9228 ∗ σ o vv + 1.987255 ∗ ε

(6)

Evaluation of Surface Soil Moisture Using Remote Sensing and Field …

227

6.6 Validation The developed model was validated using filed data collected on 1 June 2022. Figure 14 shows the relationship between actual soil moisture and soil moisture predicted by a model. Statistical metrics like R2 , RMSE, and NSE were used to evaluate the model. Soil map that was developed for the date 1 June 2022 is shown in Fig. 15. 25

Fig. 14 Relation between observed and modeled soil moisture

R² = 0.85

Observed MV

20 15 10 5 0 0

Fig. 15 Soil moisture map

5

10 15 Modelled MV

20

25

228

T. N. S. Kumar and A. A. Pathak

7 Conclusion In this work, surface soil moisture was modeled using Sentinel-1A data using an empirical equation. Sentinel-1A has demonstrated its ability to detect surface soil moisture at field scale. Semi-empirical model was developed using multiple regression analysis, with soil moisture as the dependent variable and the backscattering coefficient and dielectric constant as independent variables. On bare agricultural land, the proposed model can predict soil moisture content with an acceptable degree of accuracy, with an adjusted R2 of 0.90 at a 95% confidence level. The suggested empirical model estimates surface soil moisture content with R2 = 0.85, RMSE = 1.46, and NSE = 0.75. This study will serve as a foundation for designing irrigation water management strategies and to simulate regional hydrology based on field soil moisture. The results of the study aid local farmers in their irrigation water management. However, researchers are encouraged and recommended to do further hydrological modeling studies using field soil data and generated soil moisture maps. Declaration Myself and my team declare that the photo of soil sample collection from the agricultural field is no objection to use in the publication.

References 1. Baghdadi N, Dubois-Fernandez P, Dupuis X, Zribi M (2012) Sensitivity of main polari-metric parameters of multifrequency polarimetric SAR data to soil moisture and surface roughness over bare agricultural soils. IEEE Geosci Remote Sens Lett 10(4):731–735 2. Balenzano A, Satalino G, Lovergine F, Rinaldi M, Iacobellis V, Mastronardi N, Mattia F (2013) On the use of temporal series of L- and X-band SAR data for soil moisture retrieval. Capitanata plain case study. Eur J Remote Sens 46. https://doi.org/10.5721/EuJRS20134643 3. Beale J, Waine T, Corstanje R, Evans J (2021) Improved soil moisture estimation with Sentinel-1 for arable land at the field scale 4. Dubois PC, Van Zyl J, Engman T (1995) Measuring soil moisture with imaging radars. IEEE Trans Geosci Remote Sens 33(4):915–926 5. Esetlili MT, Kurucu Y (2016) Determination of main soil properties using synthetic aperture radar. Fresenius Environ Bull 25(1):23–36 6. Gao Q, Zribi M, Escorihuela MJ, Baghdadi N (2017) Soil moisture retrieval from Sentinel-1 and Modis synergy. In: EGU General Assembly Conference Abstracts, p 8087 7. Gururaj P, Umesh P, Shetty A (2021) Assessment of surface soil moisture from ALOS PALSAR2 in small-scale maize fields using polarimetric decomposition technique. Acta Geophys 69(2):579–588 8. Hallikainen MT, Ulaby FT, Dobson MC, El-Rayes MA, Wu LK (1985) Microwave dielectric behavior of wet soil-part 1: Empirical models and experimental observations. IEEE Trans Geosci Remote Sens 1:25–34 9. IS 2720–4 (1985) Methods for test soil. Bureau of Indian Standards. 1–39 10. Kornelsen KC, Coulibaly P (2013) Advances in soil moisture retrieval from synthetic aperture radar and hydrological applications. J Hydrol 476:460–489 11. Liang J, Liang G, Zhao Y, Zhang Y (2021) A synergic method of Sentinel-1 and Sen-tinel2 images for retrieving soil moisture content in agricultural regions. Comput Electron Agric 190:106485

Evaluation of Surface Soil Moisture Using Remote Sensing and Field …

229

12. Ma C, Li X, McCabe MF (2020) Retrieval of high-resolution soil moisture through combination of Sentinel-1 and Sentinel-2 data. Remote Sens 12(14):2303 13. Montaldo N, Fois L, Corona R (2021) Soil moisture estimates in a grass field using Sentinel-1 radar data and an assimilation approach. Remote Sens 13(16):3293 14. Parida BR, Pandey AC, Kumar R, Kumar S (2022) Surface soil moisture retrieval using Sentinel1 SAR data for crop planning in Kosi River basin of North Bihar. Agronomy 12(5):1045 15. Santos WJR, Silva BM, Oliveira GC, Volpato MML, Lima JM, Curi N, Marques JJ (2014) Soil moisture in the root zone and its relation to plant vigor assessed by remote sensing at management scale. Geoderma 221:91–95 16. Sekertekin A, Marangoz AM, Abdikan S (2020) ALOS-2 and Sentinel-1 SAR data sensitivity analysis to surface soil moisture over bare and vegetated agricultural fields. Comput Electron Agric 171:105303 17. Sekertekin A, Marangoz AM, Abdikan SAYGIN, Esetlili MT (2016) Prelim-inary results of estimating soil moisture over bare soil using full-polarimetric ALOS-2 data. The Int Arch Photogrammetry, Remote Sens Spatial Inf Sci 42:173 18. Sekertekin ¸ A, Marangoz AM, Abdikan S (2018) Soil moisture mapping using Senti-nel-1A synthetic aperture radar data. Int J Environ Geoinform 5(2):178–188 19. Srivastava PK, O’Neill P, Cosh M, Kurum M, Lang R, Joseph A (2014) Evalua-tion of dielectric mixing models for passive microwave soil moisture retrieval using data from ComRAD ground-based SMAP simulator. IEEE J Selected Topics Appl Earth Observations Remote Sens 8(9):4345–4354 20. Sutariya S, Hirapara A, Meherbanali M, Tiwari MK, Singh V, Kalubarme M (2021) Soil moisture estimation using Sentinel-1 SAR data and land sur-face temperature in Panchmahal district, Gujarat State. Int J Environ Geoinform 8(1):65–77 21. Thanabalan P, Vidhya R, Kankara RS (2021) Soil moisture estimation using RISAT-1 and SENTINEL-1 data using modified Dubois model in comparison with averaged NDVI. Geocarto Int, 1–21 22. Yang M, Wang H, Tong C, Zhu L, Deng X, Deng J, Wang K (2021) Soil moisture retrievals using multi-temporal sentinel-1 data over nagqu region of Tibetan plateau. Remote Sens 13(10):1913 23. Zayani H, Zribi M, Baghdadi N, Ayari E, Kassouk Z, Lili-Chabaane Z, Michot D, Walter C, Fouad Y (2022) Potential of C-Band Sentinel-1 data for estimating soil moisture and surface roughness in a watershed in western France 24. Zribi M, Kotti F, Lili-Chabaane Z, Baghdadi N (2012) Soil texture mapping over a semi-arid area using TERRASAR-X radar data over a semi-arid area. IEEE Trans Geosci Remote Sens Lett 9:353–357

Evaluation of the Influence of Land Use and Climate Changes in Runoff Simulation Using Semi-Distributed Hydrological Model M. S. Saranya and Vinish V. Nair

Abstract Water resources must be managed effectively to meet current and future demands, ensure sustainability, and meet the needs of a growing population. Identification of the characteristics of resources in the basin, such as the land use land cover (LULC), climatic parameters, and runoff, makes it possible to manage water resources for a long time. In this study, using the hydrological model, Soil and Water Assessment Tool (SWAT), a separation strategy, was implemented in the Meenachil River basin in the Kottayam district of Kerala to differentiate the effects of change in climate and LULC on runoff. The calibration and validation of the SWAT model were facilitated by runoff data available at gauging station Kidangoor within the study area from 1987 to 2010 for the calibration (1987–2004) and validation periods (2005–2010), respectively. Four distinct scenarios were examined to determine the relative impact of LULC and climate on runoff. The findings showed that the variation in streamflow over the study area is significantly affected by climate change (84.86%), with land use having a 15.13% influence. Keywords SWAT · Climate change · LULC change · Meenachil river basin

1 Introduction Water is a very important natural resource that is necessary for all living things to survive. Water resources must be managed effectively to meet current and future demands, ensure sustainability, and meet the needs of a growing population. With more warming in the twenty-first century, it is expected that a larger share of the world’s population will be affected by water shortages and major river floods. The hydrologic response of a watershed is strongly influenced by the changes in climatic conditions and LULC changes. These changes are accelerated by the rise in regional population, increase in greenhouse gas (GHG) emission, urbanisation, and other M. S. Saranya (B) · V. V. Nair R.I.T. Govt. Engineering College, APJ Abdul Kalam Technological University, Kerala, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_17

231

232

M. S. Saranya and V. V. Nair

anthropogenic activities within the region [1–3]. It is vital to understand how the changing conditions influence the availability of water at watershed scale. Developing countries like India are potentially at risk from the negative impacts of global climate change since they lack the resources to respond effectively to the problem. A thorough understanding of the major factors affecting runoff makes it easier to manage and plan for water resources over a long period of time. However, reliable prediction of hydrological components is a significant challenge for the research community due to the uncertainty and variability of hydrological processes in terms of location and time. Generally, experimental investigation, statistical evaluation of region-specific hydrologic parameters, and hydrological modelling are utilised to evaluate the influence of change in climate and LULC on runoff [4]. Hydrological modelling is the most widely used method for simulating hydrological processes due to its ability to depict complex water cycles and spatial variability of runoff. It is broadly acknowledged that variation in climatic parameters and LULC has the greatest influence on streamflow, which is a critical element of the hydrological cycle. Several studies have been carried out globally to assess current and future variations in streamflow by taking into account the influence of change in climate and LULC independently and together [5–9]. For the purpose of creating practical adaptation strategies and aiding decision-makers in accomplishing water security, it is essential to investigate how change in climate and LULC affect hydrological processes in a given region under different scenarios [9]. Making better planning decisions requires a quantitative estimation of how change in climate and LULC will affect regional or watershed hydrology and the identification of potential issues with water resources. This study attempted to separate the individual contributions of change in climate and LULC on runoff variation in Meenachil river basin, Kerala, which would aid in planning and management of water resources for long period in the area under investigation.

2 Materials and Methods 2.1 Study Area Meenachil river basin is a 78-km-long humid tropical river basin that begins in Vagamon hills, flows through major towns in Kottayam district, and empties into Vembanad lake. It is bordered by the basins of Muvattupuzha in the north, Periyar in the east, and Manimala in the south. It has a catchment area of 1272 km2 and is situated between within 9°25’–9°55N latitude and 76°30’–77° E longitude. The Meenachil river is threatened by extensive illegal sand mining. Historically, the riverbanks were heavily enriched with sand. Now the riverbanks are overgrown with grass and mud. Due to this illegal sand mining, the river is becoming deeper every day. Another threat to the river is sewage contamination. This river is not an outlier when it comes

Evaluation of the Influence of Land Use and Climate Changes in Runoff …

233

Fig. 1 Map of the Meenachil river basin

to monsoon floods and resulting public destruction. The location map Meenachil river basin is illustrated in Fig. 1.

2.2 Input Data Inputs for the hydrological model SWAT include climate parameters such as precipitation, maximum and minimum temperatures, relative humidity, wind speed, and solar radiation, as well as spatial data such as digital elevation model (DEM), LULC map, and soil map. The DEM of the Meenachil river basin was downloaded at 30 m resolution from NASA’s Earth Science Data System (ESDS) [10]. The digital soil map of the world at a scale of 1:5,000,000 was downloaded from the Food and Agricultural Organisation (FAO) website and used to extract the soil map for the study area. Land use and land cover maps for the study area were generated using LANDSAT satellite images with a spatial resolution of 30 m, which are freely available in the USGS’s Earth Explorer interface [11]. The LANDSAT images available for the period 2000 and 2008 that were cloud-free were chosen for this study. Landsat 7 and 5 satellite images were obtained as a tile from the USGS website using path 144 and row 53, correspondingly for 2000 and 2008. Images were then individually clipped based on the study area boundary using the ArcGIS software. With the aid of ArcGIS’s supervised classification tool and the maximum likelihood algorithm, the clipped satellite

234

M. S. Saranya and V. V. Nair

images were divided into nine different land use land cover categories (Fig. 2). These nine categories were water, grassland, tea plantation, urban area, forest, paddy field, mixed vegetation, barren land and rubber plantation. The percentage change in LULC categories for the years 2000 and 2008 is shown in Table 1. The meteorological data from 1980 to 2010 were collected from the government agencies IMD (India Meteorological Department) and IDRB (Irrigation Design and Research Board of Kerala). Data were gathered for four stations within the study area. Since the calibration and validation of a hydrological model are essential for its effective application, streamflow data from the Kidangoor station within the Meenachil river basin were collected for the purpose of calibration and validation. It was obtained from the Central Water Commission (CWC) [12] for the years 1987–2010.

Fig. 2 LULC map for the years 2000 and 2008

Table 1 Percentage area covered by LULC types between 2000 and 2008

Land type Rubber plantation

Percentage area covered (%) 2000

2008

20.897

25.265

Urban area

5.690

8.370

Paddy field

10.242

10.109

Mixed vegetation

48.238

41.337

Barren land

0.271

0.288

Grassland

1.427

1.407

Tea plantation

0.378

0.383

Water

0.630

0.627

Forest

12.226

12.214

Evaluation of the Influence of Land Use and Climate Changes in Runoff …

235

2.3 Description of the SWAT Model One of the hydrological models used by numerous researchers and agencies for simulating hydrologic components of a watershed is the SWAT model, developed in the United States [13]. It has made a significant contribution to many scientific fields worldwide. The SWAT model is a continuous-time model with a daily time step, at watershed scale [14]. The framework of SWAT model structure is shown in Fig. 3. The water balance equation governs hydrology in SWAT. It is capable of generating long-term yields in order to determine the effect of land-management practices [15]. We can simulate streamflow, sediment transport, and nutrient runoff at watershed scale using this model’s user-friendly interface. It can model the hydrologic cycle at three distinct scales, including the whole watershed, sub-watersheds, and hydrologic response units (HRUs). HRU is the model’s smallest unit, characterised by homogeneous land use, soil, and slope characteristics in a given sub-watershed based on the thresholds established by the user for each of these categories. In the current study, a threshold of 20%, 20%, and 10%, respectively, for land use, slope, and soil on area, was assigned. Depending on the user’s preference, SWAT will use either a weather generator or observed weather data for meteorological inputs. In hydrologic modelling with SWAT, the various steps include watershed delineation, HRU generation, weather definition, input table writing, hydrological simulation, calibration, and validation. A DEM-based method available in SWAT was utilised to automatically define the watershed and sub-watershed boundaries. Figure 4 displays the DEM of the study area, and Fig. 5 depicts the delineated watershed with 35 sub-basins. In order to conduct the HRU analysis, the prepared land use map was reclassified in accordance with the SWAT land use classification table. Additionally, user-defined soil types were added into the model and reclassified correspondingly. Then HRUs were obtained by overlaying the reclassified land use, soil, and slope maps of user-defined thresholds. The threshold values are established to eliminate insubstantial land use, soil type, and slope in each sub-basin, and thus preventing the formation of an excessive number of HRUs [1]. In this study area, 123 HRUs are being produced. Following that, the model was fed with all the necessary climatic variables. In the event that station data were not available, the weather generator within the SWAT can simulate weather parameters for which there are no observed data and to fill in any missing weather data. The model uses the SCS curve number method [16] to calculate runoff, and the most accurate Penman–Monteith equation was chosen to compute the potential evapotranspiration. The runoff is estimated independently for each HRU, and then it is routed in order to get the overall runoff for the watershed. The simulation included a 3-year warm-up period. The simulation was scheduled to run from 1980 to 2010, inclusive of the warm-up period.

236

M. S. Saranya and V. V. Nair

Create SWAT project

ASTER DEM (30M)

Watershed/sub-watershed delineation

Slope map

Reclassification and overlay

HRU generation

Weather definition

Soil map of the world (FAO)

Satellite image (Landsat)

Land use land cover map

Soil map

Precipitation maximum temperature minimum temperature Relative humidity Solar radiation Wind speed

Input table writing

SWAT simulation

Observed streamflow data of Kidangoor station

Calibration of SWAT Model

SWAT-CUP Software

Validation of SWAT Model

Fig. 3 Framework of SWAT model structure

2.4 Calibration and Validation Calibration is the process of selecting the best parameters for running a model, whereas validation is the process of proving that a model can make predictions that are sufficiently accurate for the intended purpose of the project. The successful application of hydrologic models relies on the calibration and sensitivity analysis of their parameters. Observed data are necessary for the calibration and validation process to be carried out effectively. The runoff data available at the Kidangoor station within the study area for the period 1897–2010 were used for the same. The SWAT model

Evaluation of the Influence of Land Use and Climate Changes in Runoff …

237

Fig. 4 DEM of Meenachil river basin

can be calibrated using the software SWAT-CUP (SWAT Calibration and Uncertainty Programs). SWAT-CUP uses the SWAT input files to change the parameter sets and run SWAT simulations. The Sequential Uncertainty Fitting algorithm (SUFI-2) was chosen from the four calibration techniques found in SWAT-CUP because it can run with multiple parameters in the fewest number of model runs to produce good prediction uncertainty ranges. 18 parameters were chosen for sensitivity analysis based on the relevant literature review and the SWAT-CUP software use manual. Two indicators, such as p-value and t-stat, govern sensitivity analysis primarily. The global sensitivity approach ranks the parameters based on their respective t-statistics and p-values after examining the sensitivity of one parameter in relation to another [14]. In the case of t-statistics, a parameter is more sensitive if its absolute value is greater, while for p-values, it is more sensitive if it is closer to zero. Calibration was performed by comparing simulated monthly runoff to observed monthly runoff data for the historical period from 1987 to 2004, and validation was performed using data from 2005 to 2010. Statistical indicators such as root mean square error to standard deviation ratio (RSR), percent bias (PBIAS), and Nash–Sutcliffe efficiency (NSE) were used to evaluate the performance of the model during the calibration and validation periods using the evaluation criteria proposed by Moriasi et al. [17] for the monthly time step.

238

M. S. Saranya and V. V. Nair

Fig. 5 Sub-basins of Meenachil river basin

2.5 Separation of the Impact of Climate and LULC Change Four scenarios were used to separate the percentage contribution of climate and LULC change on runoff. In order to ascertain the individual as well as combined contribution of climate and LULC change on runoff, simulations were carried out for the period 1983–2010 under four scenarios (scenario 1 (SR1 ), scenario 2 (SR2 ), scenario 3 (SR3 ) and scenario 4 (SR4 )) using the validated SWAT model [4]. As shown in Table 2, streamflow simulations under SR1 and SR2 were run for period 1, and simulations under SR3 and SR4 were run for period 2. The difference between SR1 and SR2 represents runoff variation caused by LULC change, whereas the difference of SR1 and SR3 indicates runoff variation caused by climate change. The difference between SR1 and SR4 shows how much LULC and climate have contributed together. The following equations were used to assess the above-mentioned analysis: ΔO L = O2 − O1

(1)

ΔOC = O3 − O1

(2)

Evaluation of the Influence of Land Use and Climate Changes in Runoff … Table 2 Scenarios considered for separation of the impact

239

Scenario

Time period

Land use map used

SR1

1983–1996

2000

Streamflow O1

SR2

1983–1996

2008

O2

SR3

1997–2010

2000

O3

SR4

1997–2010

2008

O4

ΔO = O4 − O1

(3)

ΔOm = OC + O L

(4)

ΔO = ΔOm

(5)

Theoretically,

where O1 , O2 , O3 , and O4 correspond to the average streamflow simulated under SR1 , SR2 , SR3 , and SR4 , respectively. The impact of climate change on streamflow is given by ΔOC /ΔOm × 100% and the impact of change in LULC is given by ΔOL / ΔOm × 100%.

3 Results and Discussion 3.1 Sensitivity Analysis, Calibration, and Validation Eliminating the parameters that were found to be insensitive is made easier by sensitivity analysis. The global sensitivity analysis of SUFI-2 employs the t-test and p-value, to see the effects of each parameter on the objective function [6]. We used 1000 iterations to distinguish the sensitive parameters and selected the 15 most sensitive parameters with t-statistics greater than 1 [2]. The selected parameters are SOL_ AWC, CN2, ALPHA_BNK, REVAPMN, CH_K2, ESCO, RECHARGE_DP, GW_ REVAP, SOL_BD, GWQMN, GW_DELAY, ALPHA_BF, EPCO, CH_N2 and SOL_ K, which are listed in order of sensitivity in Table 3. The table also contains the best-fitted value, p-value, and t-statistic for each of the above parameters. SOL_AWC is found to be the most sensitive parameter for the study basin. Figure 6a and b depicts the comparison of the monthly observed and simulated runoff hydrographs for the calibration (1987–2004) and validation (2005– 2010) periods, respectively. Based on the values obtained in Table 4, SWAT has a performance rating of “very good” for the calibration (NSE = 0.78 and R2 = 0.8) and validation (NSE = 0.86 and R2 = 0.9) of monthly runoff simulation. The peak flows during rainy months are under estimated, despite the fact that simulated monthly

240

M. S. Saranya and V. V. Nair

Table 3 The characteristic values of chosen sensitive parameters Serial number

Parameter and its description p-value

1

SOL_AWC (Available water 0.00 capacity of the soil layer)

t-stat 3.941

0.175

Fitted value

2

CN2 (SCS runoff curve number)

0.036

2.420

96.42

3

ALPHA_BNK (Baseflow alpha factor for bank storage)

0.062

2.121

0.325

4

REVAPMN (Threshold depth of water for revap or percolation to occur)

0.069

2.101

87.5

5

CH_K2 (Effective hydraulic 0.073 conductivity in main channel alluvium)

2.007

487.5

6

ESCO (Soil evaporation compensation factor)

0.101

−1.984

0.475

7

RECHARGE_DP (Deep aquifer percolation factor)

0.102

−1.962

0.225

8

GW_REVAP (Groundwater “revap” coefficient)

0.119

1.865

0.168

9

SOL_BD (Moist bulk density (g/cm3 ))

0.172

1.591

1.90

10

GWQMN (Threshold depth of water in the shallow aquifer required for return flow (mm))

0.185

1.437

2875

11

GW_DELAY (Groundwater delay (days))

0.198

1.248

212.50

12

ALPHA_BF (Baseflow alpha factor (days))

0.234

1.086

0.875

13

EPCO (Plant uptake compensation factor)

0.289

1.054

0.325

14

CH_N2 (Manning’s n value for the main channel)

0.312

−1.043

0.0172

15

SOL_K (Saturated hydraulic 0.326 conductivity)

−1.015

1.450

hydrographs performed satisfactorily during the calibration and validation period. This might be as a result of the streamflow simulation method’s (SCS CN method) limitations. The underestimation of extreme events can be disregarded though, as the intention of the current work is to comprehend how LULC and climate change affect runoff.

Evaluation of the Influence of Land Use and Climate Changes in Runoff …

a

241

275

Monthly streamflow (m3/s)

250 225 200 175 150 125 100 75 50 25 0 Jan-87

Dec-88

Nov-90

Oct-92

Sep-94

observed

Monthly streamflow (m3/s)

b

Aug-96

Jul-98

Jun-00

May-02

Apr-04

simulated

250 225 200 175 150 125 100 75 50 25 0 Jan-05 Jul-05 Feb-06 Aug-06 Mar-07 Sep-07 Apr-08 Nov-08 May-09 Dec-09 Jun-10 observed

simulated

Fig. 6 a Comparison of streamflow during calibration period 1987–2004. b Comparison of streamflow during validation period 2005–2010

Table 4 Model performance during calibration and validation NSE

RSR

PBIAS (%)

R2

Calibration

0.78

0.47

4.7

0.8

Validation

0.86

0.363

5.84

0.9

3.2 Relative Contribution Assessment The annual runoff simulated under the four scenarios SR1 , SR2 , SR3 , and SR4 was utilised to assess the effect climate and LULC variation. Table 5 depicts the mean annual streamflow corresponding to the four scenarios. This separation strategy was performed for the historical period 1980–2010. Due to the 3-year warm-up period considered during SWAT simulation, the actual simulation began in 1983. During period 1 (1983–1996), the simulation for scenarios SR1 and SR2 differed only by the land use map considered (see Table 2), and the simulated streamflow was 46.20 m3 /

242

M. S. Saranya and V. V. Nair

Table 5 Relative contribution rate analysis Scenario

Mean annual runoff (m3 /s)

Variation in runoff (m3 /s)

Climate contribution (%)

Land use contribution (%)

84.86

15.13

SR1

46.20

–

SR2

48.08

2.04

SR3

51.66

11.44

SR4

53.30

s and 48.08 m3 /s, respectively. The variation in streamflow caused by a change in land use was 2.04 m3 /s, while the total variation in streamflow (sum of OL and OC ) was 13.48 m3 /s. Using Eq. 4, it was determined that 15.13% of the runoff variation was due to land use change. Likewise, climate change accounted for 84.86% of the variation in runoff. Even though climate change is the most influential factor on runoff of the Meenachil basin, land use change cannot be ignored.

4 Conclusions This study employed SWAT for hydrological simulation in the Meenachil river basin. Using observed data from a stream gauge station in the basin, the simulations are effectively validated. The applicability of the model for this river basin has been proven in terms of statistical indicators such as NSE, PBIAS, RSR, and R2 during both calibration and validation. The separation of the effects of climate and LULC variation on runoff was demonstrated by simulating runoff under four distinct scenarios. When climate data from 1980 to 2010 and land use maps of 2000 and 2008 were used to separate the individual contributions on runoff simulation, it was discovered that climate change contributed 84.86% of the streamflow and LULC change contributed 15.13%. The simulations under various scenarios showed that runoff changes in the study area are primarily caused by climate change. Land use change impacts must also be considered for efficient watershed management. The management of water resources may be impacted by the various land use activities that change the hydrological cycle and the hydrological processes. The sustainable management of water resources depends on the ability to foresee the consequences of change in climate and land use on freshwater supplies [18]. The only way to identify and implement methods to improve water management actions is to examine the likely effects on a regional/watershed level.

Evaluation of the Influence of Land Use and Climate Changes in Runoff …

243

References 1. Chen Y, Niu J, Sun Y, Liu Q, Li S, Li P, Sun L, Li Q (2020) Study on streamflow response to land use change over the upper reaches of Zhanghe Reservoir in the Yangtze River basin. Geosci Lett 7:1–12. https://doi.org/10.1186/s40562-020-00155-7 2. Sinha RK, Eldho TI, Subimal G (2020) Assessing the impacts of land use/land cover and climate change on surface runoff of a humid tropical river basin in Western Ghats, India. Int J River Basin Manage. https://doi.org/10.1080/15715124.2020.1809434 3. Loukika KN, Keesara VR, Buri ES, Sridhar V (2022) Predicting the effects of land use land cover and climate change on Munneru river basin using CA-Markov and soil and water assessment tool. Sustainability 14(9):5000. https://doi.org/10.3390/su14095000 4. Guo J, Su X, Singh VP, Jin J (2016) Impacts of climate and land use/cover change on streamflow using SWAT and a separation method for the Xiying River Basin in northwestern China. Water 8:192 5. Zhang X, Srinivasan R, Hao F (2007) Predicting hydrologic response to climate change in the Luohe river basin using the SWAT model. Trans ASABE 50(3):901–910. https://doi.org/10. 13031/2013.23154 6. Zuo D, Xu Z, Yao W, Jin S, Xiao P, Ran D (2016) Assessing the effects of changes in land use and climate on runoff and sediment yields from a watershed in the Loess Plateau of China. Sci Total Environ 544:238–250. https://doi.org/10.1016/j.scitotenv.2015.11.060 7. Ahiablame L, Sinha T, Paul M, Ji J, Rajib A (2017) Streamflow response to potential land use and climate changes in the James River watershed, upper Midwest United States. J Hydrol: Regional Stud 14(1):150–166. https://doi.org/10.1016/j.ejrh.2017.11.004 8. Quansah JE, Naliaka AB, Fall S, Ankumah R, Afandi GE (2021) Assessing future impacts of climate change on streamflow within the Alabama river basin. Climate 9(55). https://doi.org/ 10.3390/cli9040055 9. Saranya MS, Vinish VN (2022) Impact evaluation and analysis at a river basin scale under projected climate and land use change. Water Supply, ws2022417. https://doi.org/10.2166/ws. 2022.417 10. Nasa EARTH DATA Homepage. https://search.earthdata.nasa.gov/search. Accessed 18 July 2021 11. USGS Earth Explorer Homepage. https://earthexplorer.usgs.gov/. Accessed 19 July 2021 12. India WRIS Homepage. https://indiawris.gov.in/wris/#/. Accessed 11 May 2020 13. Zhao K, Li J, Ma X, Deng C (2022) The effects of land-use and climatic changes on the hydrological environment in the Qinling mountains of Shaanxi province. Forests 13:1776. https://doi.org/10.3390/f13111776 14. Abbaspour KC (2015) SWAT-CUP: SWAT calibration and uncertainty programs—a user manual. Eawag: Swiss Federal Institute of Aquatic Science and Technology, p 100 15. Arnold JG, Moriasi DN, Gassman PW, Abbaspour KC, White MJ, Srinivasan R, Santhi C, Harmel RD, Van Griensven A, Van Liew MW, Kannan N, Jha MK (2012) SWAT:model use, calibration, and validation. Trans ASABE 55(4):1491–1508. https://doi.org/10.13031/2013. 42259 16. Neitsch SL, Arnold JG, Kiniry JR, William JR (2009) Soil and water assessment tool theoretical documentation. AgriLIFE Research & Extension 17. Moriasi DN, Arnold JG, Van Liew MW, Bingner RL, Harmel RD, Veith TL (2007) Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans ASABE 50(3): 885–900. https://doi.org/10.13031/2013.23153 18. Raihan F, Ondrasek G, Islam MS, Maina JM, Beaumont LJ (2021) Combined impacts of climate and land use changes on long-term streamflow in the Upper Halda Basin, Bangladesh. Sustainability 13(21):12067. https://doi.org/10.3390/su132112067

Flood Damage Assessment of a River Basin Using HEC-GeoRAS K. C. Amal Vishnu and Vinish V. Nair

Abstract Flood is one of the most catastrophic events among the many different types of natural hazards. It seriously harms people, property and places used for industry and agriculture. A GIS extension called HEC-GeoRAS provides a series of steps, tools and options for preparing river geometry grid data for input into HECRAS, which is utilised to create the final inundation map. The DEM and a Basin Landuse Map must be provided as input data for the preparation of river geometry using HEC-GeoRAS model. The flood hydrograph is derived from the synthetic unit hydrograph and probable maximum precipitation, and the PMP value obtained was about 157.3 cm. The total area of the basin was found to be 787 km2 in which the inundated area was 274.279 km2 , which is about 35.851% of the total area. Keywords DEM · Landuse map · HEC-GeoRAS

1 Introduction 1.1 General Background Among the many different types of natural hazards, flooding is one of the most catastrophic occurrences. It seriously harms people, infrastructure and places used for industry and agriculture. The most frequent natural calamities in a tropical nation like India are floods. Flooding is defined by the European Union Floods Directive (EUFD) as the water flooding of territory that is not normally covered by water. Thus, the land that is normally dry is submerged by an overflow of water. When water overflows from a river, lake, or ocean, the levees may be breached or overtopped. K. C. A. Vishnu (B) · V. V. Nair Department of Civil Engineering, Rajiv Gandhi Institute of Technology Kottayam, APJ Abdul Kalam Technological University, Kerala, India e-mail: [email protected] V. V. Nair e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_18

245

246

K. C. A. Vishnu and V. V. Nair

Flooding can also happen when precipitation collects on soggy terrain. Thus, the primary forces of flood are heavy rainfall, drainage cramming, impervious soil and erosion of banks. Although the size of lakes, rivers and other water bodies will alter due to seasonal differences in rainfall and snowmelt, these changes are unlikely to be significant unless they result in property flooding or domestic animals drowning. Some floods form gradually, while others, like flash floods, can form quickly and without any outward signs of rain. Also, floods can either be very massive, affecting entire river basins, or very minor, affecting a specific area or town. Due to the continuous developmental activities, it is difficult or rather impossible to prevent flooding of the urban areas. Hence, flood management aiming at reduction of loss of life and property must be the prime concern. Many studies in the area of developing models for the forecasting of rural and urban flooding have been conducted in recent years. Every year, in most parts of India, a repeated occurrence of flood is found with states in southern part of India the most affected by its ferocity. To identify the flood-prone areas and to have knowledge on probability of occurrence or return period of floods has become inevitable for planning and mitigation of damage caused by this natural phenomenon. To understand the effect of flooding on life and property, flood inundation mapping (FIM) is done [8]. FIM gives the significant data such as spatial extent and depth of the areas affected by flood, which are required to adopt appropriate flood management strategies. Spatial analysis tools such as geographic information system (GIS) and remote sensing (RS) are effectively used for assessment of the flood risk. In spite of the fact that floods cannot be prevented, studies done to date show, the inundation caused due to flood can be reduced if the flood-prone areas were known in advance this can be achieved with the help of a flood inundation map developed for the study area using available technological advances. It also helps in effective and complete flood management system development.

1.2 Objectives The objectives of the study are • To obtain the water delineation of the digital elevation model of Vamanapuram basin using ArcHydro tool. • To determine the synthetic unit hydrograph and hence to develop a flow hydrograph for Vamanapuram basin. • To prepare the probable maximum precipitation. • To prepare flood inundation map using HEC-RAS software. • To conduct a flood damage assessment of Vamanapuram basin using HECGeoRAS software.

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

247

1.3 Study Area Study area comprises the region from Chellangi (Panavoor) to Thuruthu (Chirayinkeezhu) of Vamanapuram River Basin. It is spread over Thiruvananthapuram and Kollam districts of Kerala state. The basin is bounded by Kottarakara Taluk of Kollam district in the north and by the Arabian Sea in the west. The Vamanapuram river rises in the Western Ghats’ Chemmunji mottai at a height of roughly 1706 m above sea level and empties into Anjengo Lake. The river has a drainage area of 787 sq. km and is 80 km long. It lies between 51°18' 59'' E and 51°26' 58'' E longitude and 33°38' 6'' N and 33°46' 19'' N latitude.

2 Literature Review 2.1 General Flood, being a disastrous natural phenomenon, has to be assessed and managed so as to reduce its dreadful effects. Proper and timely warning if provided can mitigate the flood damages to a greater extent. Developmental activities are continuous in urban regions and thus to prevent urban flooding it is almost become impossible. In the past few decades, various research has been published in connection with the development of models for the forecasting and prediction of floods in both urban and rural settings. Floods are a common occurrence in most of India each year, but they are most severe in the southern states of India.

2.2 Flood Damage Assessment To understand the effects mainly on an important structure or on any particular region flood inundation mapping or FIM is required. FIM gives the significant data such as spatial extent and depth of the areas affected by flood, which are required to adopt appropriate flood management strategies. Sahoo and Sreeja [8] conducted a study in an urban catchment in northeast India to quantify the flood risk and to use flood inundation mapping (FIM) as an effective tool in flood management. Storm water management model (SWMM) was used to estimate the surface runoff. The flooding pattern for peak rainfall intensities was then acquired using the local drainage network and correlating to various return periods. Based on the potential flood hazard brought on by the 100-year return period flood, a flood hazard map was created. Based on the hazard rankings developed using the shared impact of vulnerability parameters such as flood depth, inundated area, land use, affected population, and roadways, flood risk zones have been defined. The effectiveness of HR for locating the worst flood-affected area is discussed in the study.

248

K. C. A. Vishnu and V. V. Nair

Getahun and Gebre [3] carried out a flood hazard assessment and mapping of the flood inundation area in Awash River basin, Ethiopia. Factors causing floods such as slope, elevation, rainfall, drainage density, land use and soil type were rated and the flood hazard regions are delineated using a multi-criteria evaluation technique in a GIS. To obtain the flood danger map, a final weighted overlay analysis of all elements was obtained by calculating the weight of each flood-generating factor by pair-wise comparison. All return periods had high flood levels, especially from Dubti to Lake Abe. According to the study, there were less inundated areas in the upper and intermediate reaches of the Awash River Basin than there were in the lower reaches. The Awash River Basin’s northwest, southwest and western escarpments all fall under the low to very low flood hazard danger category. In particular in lowlying flood-prone areas, afforestation and good land use management are crucial to reducing the detrimental consequences of floods. The study is very beneficial for the relevant bodies in helping them create measures based on the potential for flooding in the area. The findings of this paper are crucial for planning land use, formulating policies, making investment decisions and for security considerations. Kourgialas and Karatzas [6] presented a method to evaluate the benefits of a flood warning system, as well as a method to determine the hazardous areas. A sound flood control strategy and an evaluation of the flood-prone areas make up a trustworthy flood management plan. They analysed six factors in order to determine the spatial distribution of the dangerous locations—flow accumulation, rainfall intensity, slope, geology, land use and elevation. Five regions were chosen for the study that ranged in flood threat from extremely low to very high. The locations with a high risk of flooding are delineated on the created map of flood-hazard areas. The Koiliaris River basin was examined using the suggested methodology to identify the settlements and areas at risk of floods. The results were then verified using information from previous floods in the basin. Chithra and Sumam [2] used hydraulic model HEC-RAS for flood inundation mapping of Kurumali river basin for probable maximum flood event. The model’s inputs were created with the help of ArcGIS, this is a software product developed by ESRI, which is a compatible programme Hydrologic Engineering Center’s Geographical River Analysis System (HEC-GeoRAS). The flow and stage hydrographs for the research region could be reasonably predicted by the hydraulic model. Hydrologic Engineering Center’s River Analysis System (HEC-RAS) was used to route the flood for the greatest likely flood occurrence. The mapping of flood inundation was done using the results that were collected. For studies on flood risk assessment, the created flood inundation maps could be used. Accuracy of model predictions can be increased by using high-resolution Digital Elevation Model (DEM) or field observed data.

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

249

3 Methodology 3.1 General The generalised methodology includes watershed delineation from the DEM developing the cross-section and streamline in HEC-GeoRAS, development of flood inundation map and assessment of the flood impact of Vamanapuram river basin using HEC-GeoRAS software.

3.2 Description of the HEC-GeoRAS Model The final inundation map is produced using the GIS extension HEC-GeoRAS, which has a number of steps, tools and options for preparing river geometry GIS data for import into HEC-RAS. Triangular Irregular Network (TIN), DEM and land use are the input data needed to prepare the river geometry for the HEC-GeoRAS model. The river geometry file and stream flow data are also utilised as input files for the HEC-RAS model, which generates the river’s water surface level. HEC-GeoRAS serves as an interface for data management between ArcGIS and HEC-RAS. The river streamline, centreline, banklines, flow route centrelines and XS cut lines must be digitally recovered from an earlier river file, from aerial photographs or from topographical data using the HEC-GeoRAS interface. The HEC-GeoRAS geodatabase file contains information on river reach, cross-section and other relevant topics. With the exception of the river and reach names, which are manually filled in for the cross-section and river data layers, all other properties are automatically computed by the HEC-GeoRAS. A GIS environment will be used to store the outcomes of the HEC-RAS model simulation, and the HEC-GeoRAS tool will be used to conduct additional studies. The GIS data transmitted between HEC-RAS and ArcGIS are in.sdf file format. The HEC-RAS editing tools can be used to modify the exported GIS geometric data in the model. The Geometric Editor, a Graphical User Interface (GUI) used to handle the geographic data, imports the.xml file exported from the HEC-GeoRAS [3]. For the cross-sections of each reach, the Manning friction values are entered in this editor. The steady flow data editor receives the stream flow data. The data for the reaches are taken out of the geometric editor by this editor. The model needs to be aware of the initial water level at the beginning and finish of unconnected reaches as well as at junctions with other reaches in order to calculate the water surface level (boundary conditions). Water surface profile changes for various flow rates with variable recurrence intervals in desired river lengths, current velocity values, normal depth, critical depth, hydraulic features and parameters in the river are the output data of the HEC-RAS model [1].

250

K. C. A. Vishnu and V. V. Nair

3.3 Watershed Delineation of Vamanapuram River Basin The region upstream from a given outflow point serves as the physical boundary for watersheds or catchments. Both physically using paper maps and digitally using a GIS environment are acceptable methods. In this study, watershed delineation is done using the DEM obtained from Bhuvan (National Remote Sensing Centre, Cartosat-1 A image).

3.4 Preparation of Flood Inundation Map The flood inundation mapping of Vamanapuram River basin is done using HECRAS, for which a set of data such as cross-section, streamline and centreline of the river and flood hydrograph are required. The cross-section, streamline and centreline are extracted with the help of HEC-GeoRAS. The flood hydrograph is developed as per the procedure proposed by Central Water Commission (CWC) [10].

3.4.1

Estimation of Flood Hydrograph

Flood hydrograph estimation requires a Synthetic Unit Hydrograph (SUH) and the Probable Maximum Precipitation (PMP) for duration equal to the base period of SUH [4]. The step wise procedure of estimation of SUH and PMP is as follows: (i)(i) Development of SUH A catchment’s SUH is a unit hydrograph with a unit duration. It was created using relationships between the physiographic and unit hydrograph parameters of sample catchments in a hydro-meteorologically uniform region. Physiographic characteristics of the region include catchment area, length of main stream and equivalent stream slope. Catchment area (A) is considered as the area enclosed within the watershed boundary. Length of main stream (L) as well as the length from centre of gravity to the gauging site (Lc ) is obtained from the DEM. Equivalent stream slope (m/km) is obtained from the Eq. 1 [10]. s=

Σ(Li (Di−1 + Di ) L2

(1)

where, Li = length of the ith segment (km) Di-1 , Di = depth of the river at the point of intersection of (i-1) and ith contours from the datum at the level of point of study (m) L = length of longest stream (km) Catchment plan, showing the foresaid measurement, is shown in Fig. 1. The various parameters of the 1-h SUH are derived using the following Eqs. (2–10) [10].

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

) ( L ∗ L c 0.405 t p = 0.553 √ s qp =

2.043 t p0.872

251

(2) (3)

Q p = qp ∗ A

(4)

TB = 5.083 ∗ t p0.733

(5)

Tm = t p + 0.5tr

(6)

where, tp = basin lag qp = peak discharge per sq.km area of catchment Qp = peak discharge of the unit hydrograph for the catchment area (A) TB = time base of SUH Tm = peak time of SUH Width of SUH at 50 and 75% of peak discharge ordinate Qp in hours is computed using Eqs. 7–10. W50 = 2.197/q1.067 n

Fig. 1 1-h SUH

(7)

252

K. C. A. Vishnu and V. V. Nair

W75 = 1.325/q1.088 n

(8)

WR50 = 0.799/q1.138 n

(9)

WR75 = 0.536/q1.109 n

(10)

A pictorial representation of a sample SUH is shown in Fig. 1. (ii)(ii) Estimation of PMP It is defined as the greatest depth of precipitation, which is possible for a given time and duration over a given storm area under known meteorological conditions [6]. There are in general two approaches for the estimation of PMP: • Traditional frequency analysis method based on Chow’s general frequency equation. • Approach specified by World Meteorological Organisation (WMO), 1986, which is basically frequency analysis method. Since the latter involves graphical methods, the approximation of values varies from person to person. Hence, the traditional frequency analysis is used. It was proposed by Chow [6]. Chow’s equation is given in Eq. 11. P M P = Xn + K m σn

(11)

where, Xn = mean of n annual maximum values σn = standard deviation Km = frequency factor Km =

K m − K n−m Σ K n−m

(12)

Xm = highest value in the series Xn-m = mean of the series excluding the highest value ΣXn-m = standard deviation excluding the highest value The PMP, thus, obtained is divided into corresponding interval by multiplying its value with the 48-h time duration coefficient and hence the storm rainfall is obtained. The 48-h time duration coefficient is obtained from the Central Water Commission (CWC), Flood Estimation Report for Kaveri basin [10]. The rainfall increment values are then computed. Then the effective rainfall values are obtained by reducing the loss rate, i.e., 0.1 cm/hr (CWC). (iii) Derivation of flow hydrograph

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

253

The rainfall excess increments are arranged against the hourly SUH ordinates such that the maximum value of rainfall excess comes against the peak discharge of SUH, the next lower value of rainfall excess increment comes against the next lower discharge ordinate and so on. The summation of the product of unit hydrograph ordinate and 1-h rainfall excess produces the direct runoff [5]. The design base flow rate is 0.05 cumec per sq.km [10], and hence the total base flow for area of the catchment is obtained by multiplying the catchment area with the design base flow rate and the base flow is finally derived [9]. The base flow is multiplied to the total direct runoff values to get the maximum discharge. The maximum discharge ordinates hence obtained is plotted against the duration to obtain the flood hydrograph.

3.4.2

Extraction of Cross-Section, Streamline and Centreline of the River

The watershed delineation done using ArcHydro tool is used to extract the crosssection, streamline and centreline of the river. The extraction is done using HECGeoRAS, which is a customised toolbar in ArcGIS; the RAS geometry menu being mainly used [3]. The polylines (XS cutlines) are created at the points were the cross-section are needed. The cross-section, centreline, bank line are generated by developing corresponding feature class [3]. This acts as a pre-processing input for HEC-RAS one-dimensional model used for comparison with the field data.

3.4.3

Development of Flood Inundation Map

The flood hydrograph and the river geometrical data derived are used to develop the flood inundation details about that particular river cross-section by running the unsteady flow. condition and hence the rise in water level corresponding to the cross section is obtained. The RAS GIS import file (final river geometry file), which can be used as input for HEC-RAS, is produced by the HEC-GeoRAS model following the creation of the RAS geometry data. The cross-section is then verified, the river geometry is edited and the river geometry file in the HEC-RAS model is finalised and corrected. Boundary conditions are used following the creation of the final river geometry file [3]. The boundary conditions for the upstream generally have a flow or stage hydrograph, and the downstream end has a typical rating or normal depth, internal boundary conditions can be added to define a gate operation or to add a flow within a river system. The unsteady flow simulation is done and the water level for different return periods is generated in HEC-RAS for a particular cross-section. In order to get the flood inundation map of the basin, 2D modelling is to be done [1]. In 2D modelling, a 2D mesh is created over the terrain of the basin, the terrain is the float file of the DEM, which is imported into HEC-RAS from the ArcGIS, the boundary conditions with the main river, tributaries and lower boundary are also set. Unsteady conditions are applied to the boundary conditions with the respective

254

K. C. A. Vishnu and V. V. Nair

flood hydrograph, which was obtained from the synthetic unit hydrograph that was generated for the basin using the PMP. The inundation map for each return period is then exported into HEC-GeoRAS for final inundation area mapping along the river.

3.5 Flood Damage Assessment The inundated area is exported in TIF file format and is imported on the HECGeoRAS. The land use land cover map of the corresponding river basin is prepared using Arc GIS 10.1 software. It is then used as a layer on which the inundated map is to be overlaid. The land use land cover map may contain attributes such as plantation, builtup area, forest, water body, etc. of the corresponding basin and hence the percentage of inundation over the attributes is known from the attribute table. This helps one to assess the possible damage caused by a flood.

4 Results and Discussion 4.1 General Daily rainfall data of the past 21 years were collected from Indian Meteorological Department (IMD), Trivandrum. Toposheets of the Vamanapuram basin were collected from Survey of India Department. Watershed was delineated from DEM using HEC-GeoRAS, for which the DEM was obtained from Bhuvan (National Remote Sensing Centre). The process of developing the cross-section and streamline was carried out in HEC-GeoRAS, whereas the development of flood inundation map was done in HEC-RAS. The assessment of the flood impact of Vamanapuram river basin using HEC-GeoRAS software was then carried out.

4.2 Watershed Delineation of Vamanapuram River Basin Using the DEM acquired from Bhuvan, watersheds or catchments were defined as the region upstream from a specific outlet point, (National Remote Sensing Centre, Cartosat-1 A image). Watershed delineation was done in Arc GIS, and the resultant delineated map is shown in Fig. 2.

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

255

Fig. 2 River delineation map

4.3 Preparation of Flood Inundation Map The flood inundation mapping of Vamanapuram River basin was done using HECRAS. The required cross-section, streamline and centreline of the river were extracted using HEC-GeoRAS. The flood hydrograph was developed as per the procedure proposed by Central Water Commission (CWC) [10].

4.3.1

Estimation of Flood Hydrograph

(i) Development of SUH Catchment area (A) was taken as the area enclosed within the watershed boundary and is equal to 787 km2 . Length of main stream (L) is about 88 km, and the length from centre of gravity to the gauging site (Lc ) was obtained from the DEM as 45 km. Equivalent stream slope (m/km) was obtained using the Eq. 1 as 0.411 m/km. The various parameters of the 1-h SUH derived using the Eqs. 2–10 are as follows: t p = 18.966 hr qp = 0.1569 cumecs Qp = 123.655 cumecs

256 Table 1 Synthetic unit hydrograph ordinates

K. C. A. Vishnu and V. V. Nair

Duration (hr)

Discharge (cumec)

0

0

4

7.5

8

22.5

12

53.7

16

110

20

122.5

24

101.3

28

62.5

32

35

36

17.5

40

7.5

44

0

48

0

T B = 43.942 hr T m = 19.466 hr W50 = 15.852 h r W75 = 9.939 hr WR50 = 6.575 hr r WR75 = 4.18 hr The developed SUH coordinates are given in Table 1, and SUH graph is shown in Fig. 4.5. (i) Estimation of PMP The PMP is referred to as the maximum depth of precipitation that is feasible across a specific storm area for a given amount of time and duration under established meteorological circumstances. The traditional frequency analysis as proposed by Chow (Eq. 11) is used, and the corresponding parameters are obtained using Eqs. 11 and 12 and are as follows Xm = 1671 cm Xn − m = 1518 cm ΣXn − m = 219 cm Xn = 1408 cm sn = 239 cm Km = 0.69 PMP = 157.3 cm

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

257

Table 2 Computed effective incremental rainfall Duration

TD coefficients

Storm rainfall (cm)

Rainfall increment (cm)

Effective rainfall (cm)

0

0

0

0

0

4

0.3

50

50

50

8

0.5

74

24

23

12

0.6

88

14

14

16

0.7

102

14

14

20

0.7

110

7.9

7.8

24

0.8

120

9.4

9.3

28

0.8

126

6.3

6.2

32

0.9

137

11

11

36

0.9

142

4.7

4.6

40

0.9

146

4.7

4.6

44

1

153

6.3

6.2

48

1

157

4.7

4.6

The Probable maximum precipitation thus obtained is equal to 157.3 cm. It was then divided into corresponding interval by multiplying the PMP value with the 48-h time duration coefficient and hence the storm rainfall was obtained. The computed rainfall increment values and the effective rainfall values are given in Table 2. (ii) Derivation of flow hydrograph The base flow of a catchment is obtained by multiplying the design base flow rate with the area of the catchment. The design flow rate as per CWC from flood estimation report Kaveri basin report is 0.005 cumec per sq.km, and hence the total base flow for the catchment of area 788 km2 was obtained as 39.4 cumec. To obtain the maximum discharge, the entire direct runoff values were added to the base flow. The maximum discharge ordinates as given in Table 3 are plotted against the duration to obtain the flood hydrograph and are shown in Fig. 3.

4.3.2

2D Modelling in HEC-RAS for River Catchment in Vamanapuram Basin

The 2D modelling is done by creating a 2D mesh of the required space in the loaded terrain of Vamanapuram basin. Mesh created had 134,048 cells. The main river, the tributary within the mesh were marked using the tools and a lower boundary is also given, to mark the final flood area. The flood hydrograph generated is given as an input for the main river and for tributaries, and the normal depth value of 0.01 is given as the input for the lower boundary. The unsteady flow analysis is done respective to the above conditions and

258

K. C. A. Vishnu and V. V. Nair

Table 3 Flood hydrograph ordinates Direct runoff (cumec)

Base flow (cumec)

0

39.4

Flood hydrograph ordinate (cumec) 39.4

34.6425

39.4

74.0425

174.7125

39.4

214.1125

585.9207

39.4

625.3207

2584.45

39.4

2623.85

6153.91

39.4

6193.31

1423.9741

39.4

1463.3741

878.5625

39.4

917.9625

326.83

39.4

366.23

108.36

39.4

147.76 85.84

46.44

39.4

0

39.4

39.4

0

39.4

39.4

Flood hydrograph ordinate (cumecs)

Flood hydrograph

Fig. 3 Flood hydrograph

hence the flood inundation map over the basin is obtained, which is shown in Fig. 6. Apart from the Inundation map, the arrival time and the water surface elevation were also obtained as shown in Figs. 4 and 5.

4.3.3

Flood Damage Assessment in HEC-GeoRAS

The landuse map was obtained from the Landuse Board with major classifications. The flood inundation map obtained was overlayed on the landuse map in order to assess the damage caused over the classification of the landuse. Figure 7 shows the image with the landuse map overlayed with inundation map.

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

259

Fig. 4 Arrival time of the flood over the terrain

Fig. 5 Maximum water surface elevation of Vamanapuram basin

The main classification contains the water body in that region, Plantation 1 which represent the plantations done on high elevated area example Rubber, Plantation 2 that represent the plantations done on the low-lying areas such as wheat, paddy etc., the other classifications are mixed vegetation, forest, builtup area and barren land.

260

K. C. A. Vishnu and V. V. Nair

Fig. 6 The flood inundation map of Vamanapuram basin

Fig. 7 Flood inundation map over landuse map

The damage area affected on each classification is individually calculated by clipping the flood inundation over the landuse classification (Table 4). The area of the classification water body is 32.12 km2 , and hence the total area of the catchment becomes 787 km2 . Therefore, the total flood damage of the basin thus obtained is 260.383 km2 and is 33.085% of the total area.

Flood Damage Assessment of a River Basin Using HEC-GeoRAS

261

Table 4 Classification-wise damage assessment Classifications of the landuse map

Total area of the classification on the land use map (km2 )

Clipped area excluding the flood-inundated region (km2 )

Total damage area of the classification (km2 )

Percentage of damage in each classification (%)

Plantation 1

321.802

174.8031

146.998

Plantation 2

94.662

65.2983

29.3637

31.019

Mixed vegetation

205.994

128.17004

77.82396

37.779

Forest

122.249

45.679

119.7361

2.5129

2.055

Builtup

4.6201

3.5279

1.0922

23.640

Barren land

5.5532

2.9610

2.5922

46.679

5 Conclusion The water delineation for the Vamanapuram basin was obtained using ArcHydro tool from the digital elevation model. 21 years of daily rainfall value of the Vamanapuram region was collected from Indian Meteorological Department. Using 1-h synthetic unit hydrograph and the probable maximum precipitation values, the flow hydrograph was developed. The Vamanapuram basin terrain was then used to create the flood inundation map using the HEC-RAS programme. The flood damage assessment of Vamanpuram basin was done in HEC-GeoRAS and ArcGIS. The PMP value obtained was 157.3 cm and a 48-h duration flood hydrograph was derived. The total area was found to be 787 km2 in which the inundated area was about 260.383 km2 and was 33.085% of the total area. The landuse classification that affected the most was barren land, about 46.679% of the classification was affected due to the flood. The remaining classifications such as Plantation 1, Plantation 2, mixed vegetation, forest and builtup area in which the area affected are 45.679%, 31.019%, 37.779%, 13.423%, 2.055% and 46.679%, respectively.

References 1. Ali H, Lee Teang S, Majid M, Hadi M (2012) Incorporation of GIS based program into hydraulic model for water level modeling on river basin. J Water Resour Protect 2. Chithra M, Sumam KS (2014) Flood risk assessment by GIS and hydraulic model. In: Proceedings of international conference on materials for the future—innovative materials, processes, products and applications—ICMF, pp 315–318 3. Getahun AA, Gebre SL (2015) Flood hazard assessment and mapping of flood inundation area of the Awash River Basin in Ethiopia using GIS and HEC-GeoRAS/HEC-RAS. J Civil Environ Eng 5(4):1–12 4. Singh KJ, Sarkar S (2012) Development of GIUH for the catchment contributing to Loktak lake, North East India. J Ind Soc Remote Sens 5. Reddy NA, Seelam JK, Rao S, Nagaraj MK (2018) Flood estimation at ungauged catchments of western catchments of Karnataka, West coast of India. ISH J Hydraul Eng

262

K. C. A. Vishnu and V. V. Nair

6. Kourgialas NN, Karatzas GP (2011) Flood management and a GIS modelling method to assess flood-hazard areas—a case study. Hydrol Sci J 56(2):211–225 7. Lan P, Lin B, Zhang Y, Chen H (2017) Probable maximum precipitation estimation using the revised km-value method in Hong Kong. J Hydrol Eng 8. Sanat Nalini Sahoo and, Pekkat, Sreeja (2017) Development of flood inundation maps and quantification of flood risk in an urban catchment of Brahmaputra River. J Risk Uncertainity Eng Syst 20:1–11 9. Tazyeen S, Nyamathi SJ (2015) Flood routing in the catchment of urbanized lakes. Aquatic Procedia 10. Flood Estimation Report for Kavery Basin Subzone – 3 (i), (1985) Central Water Commission

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques Sampath Kumar, Talari Reshma, Savitha Chirasmayee, Kasa Priyanka, Kokku Priyanka, and Gokla Ram

Abstract Floods pose a significant natural hazard that can cause extensive damage to infrastructure, property, and human life. The use of geospatial techniques, such as remote sensing, Global Positioning Systems (GPS), and Geographic Information System (GIS), has become increasingly important for flood hazard mapping. The research paper aims to develop flood hazard maps for the Amaravati region, the new capital city of Andhra Pradesh, India, using geospatial techniques, including an Analytic Hierarchy Process (AHP). The region is susceptible to floods due to its location and topography, which includes low-lying areas and water bodies. In the present study, remote sensing techniques were employed to extract information on land use, land cover, topography, and drainage patterns, and GIS to integrate this information with other relevant data, such as rainfall and river flow data, to develop flood hazard maps. Additionally, the study utilized the AHP process, a multicriteria decision-making method, to weigh and rank various factors involved in the development of the d hazard maps. The AHP process provided a structured and systematic approach to prioritize the importance of different variables and factors in the flood hazard mapping process. The developed maps provide valuable information for decision-makers, urban planners, and emergency management agencies to plan for and mitigate the impact of potential flood events in the region. The AHP process, in combination with geospatial techniques, contributed to the accuracy and reliability of the flood hazard maps. By utilizing geospatial techniques and the AHP, this research paper contributes to the existing knowledge on flood hazard mapping. The findings provide valuable insights that can be applied in flood risk management and disaster preparedness in the Amaravati region. Keywords Analytic hierarchy process · Amaravati · Flood hazard maps · Geographic information systems · Remote sensing

S. Kumar · T. Reshma · S. Chirasmayee (B) · K. Priyanka · K. Priyanka · G. Ram National Institute of Technology Andhra Pradesh, Tadepalligudem, India e-mail: [email protected] T. Reshma e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_19

263

264

S. Kumar et al.

1 Introduction Floods are one of the most devastating natural disasters that cause widespread damage to infrastructure, property, and human life worldwide. With the increasing frequency and severity of floods, accurate flood hazard mapping has become crucial for mitigating and managing the impact of these disasters. Flood hazard mapping involves the identification and assessment of areas at risk of flooding based on factors such as topography, land use, and hydrological data. Geospatial has emerged as powerful tools for flood hazard mapping. Remote sensing images provide high-resolution data on land use, land cover, and topography, while Geographic Information systems (GIS) enables the integration and analysis of multiple datasets, including rainfall and river flow data. The integration of these technologies has resulted in more accurate and reliable flood hazard maps. Several studies have highlighted the significance of geospatial techniques for flood hazard mapping. Flood hazard zonation mapping using GIS emphasized its effectiveness in identifying flood-prone areas is reviewed [1]. Few studies [2, 3] assessed flood hazard using remote sensing and GIS and concluded that remote sensing techniques provide accurate data for flood hazard mapping. Reference [4] conducted a case study on flood hazard mapping using remote sensing and GIS and machine learning in Yuyao, China and demonstrated its usefulness in identifying high-risk areas. Few studies [5] developed novel flood hazard maps using remote sensing and GIS in the Mahanadi River Basin, India and found that it was effective in predicting potential flood-prone areas. Sharma et al. [6] developed flood hazard maps using remote sensing and GIS in the Alaknanda River Basin, India and showed that the technology can provide valuable insights for flood risk management. Moreover, several studies have demonstrated the effectiveness of the Analytic Hierarchy Process (AHP) for flood hazard mapping. The AHP process was used to develop a flood hazard map for the River Basin in Philippines [7]. The authors found that AHP helped to identify the most critical factors affecting flood risk and enabled the creation of a more accurate flood hazard map. References [8, 9] employed AHP to develop a flood hazard map for the different river basins, India. The study demonstrated that AHP can be used to assess and rank different factors affecting flood risk, resulting in a more reliable flood hazard map. Furthermore, the integration of AHP with other geospatial techniques, such as remote sensing and GIS, has resulted in more accurate and reliable flood hazard maps. A recent study by Gupta and Dixit [10] used AHP to develop a flood hazard map for the Assam region, India and found that the integration of AHP with geospatial techniques resulted in a more comprehensive understanding of flood risk. Flood hazard mapping is essential in areas susceptible to floods, such as the Amaravati region of Andhra Pradesh, India which is the capital of the state. Geospatial techniques can aid in flood hazard mapping by providing high-resolution data on land use, topography, and hydrological characteristics. The integration of these geospatial techniques with AHP can enhance the accuracy of flood hazard maps by ranking and assessing the relative importance of various factors affecting flood risk. In this paper,

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques

265

we aim to develop a flood hazard map for the Amaravati region using geospatial techniques and AHP. The map will serve as a valuable tool for flood risk management, disaster preparedness, and infrastructure planning.

2 Study Area The study area of this research is the Amaravati region, located in the state of Andhra Pradesh in India, between the coordinates of 16.58° N, 16.29°S, 80.67° E, 80.16°W, as depicted in Fig. 1. It covers an area of 1020.632 square kilometers, encompassing the regions of Tadikonda, Pedakurapadu, Thullur, Vaddeswaram in Guntur district, and the outskirts of Amaravati, Satenapalle, Guntur, Mangalagiri in Guntur district, and Vijayawada in Krishna district, as shown in Fig. 1. Amaravati is situated between the Krishna River and the low hills of the Deccan Plateau, and it is the point where the Kondaveeti Vagu meets the Krishna River, causing inundation of 13,500 acres of land during the monsoon season each year. Amaravati is known for its rich cultural heritage as an ancient Buddhist center, and it is a popular destination for sightseeing and pilgrimage. From various studies, it is observed that 70% of the Amaravati area is vulnerable to flooding, and the region experiences floods three times on average every year, with severe flooding occurring in 1977, 1998, 2009, and 2019 in the Krishna River. The region also experiences frequent water shortages and droughts. These observations demonstrate the need for effective flood management and mitigation strategies to minimize the adverse impacts of floods on the region’s socio-economic and environmental well-being.

Fig. 1 Location of Amaravati study area

266

S. Kumar et al.

3 Methods and Methodology 3.1 Data Used In this study, we obtained essential data from various sources to analyze the watershed characteristics and develop effective management strategies for water resources. Specifically, we collected and downloaded Digital Elevation Model (DEM) data from the USGS Earth Explorer website, Landsat 8 data for generating the Land Use and Land Cover (LULC) map, and soil data from the FAO website as shown in Table 1 for the year 2019. We utilized flow direction data as input for preparing the Flow Accumulation Map, and we generated the slope map and rainfall intensity map using geospatial technologies. The seven maps produced using these geospatial technologies are essential inputs for understanding the hydrological processes and identifying the areas that are prone to floods and water scarcity in the region are shown in Fig. 2a–i. These maps serve as a critical resource for policymakers and stakeholders to develop sustainable water resource management strategies for the region.

3.2 Methodology In the present study, AHP a widely used multi-criteria decision-making method for analyzing complex decisions is used for preparing flood hazard maps. It is used for ranking a set of alternatives or selecting the best alternative among the given alternatives. In this study, using AHP, the flood hazard map of the Krishna River basin is obtained by choosing six different parameters that influence flood occurrence and assigning weightage to each parameter as shown in Table 2 and Fig. 3. The AHP method involves several steps. Firstly, the problem and criteria are defined. In this study, the problem is to estimate the flood hazard map of the Krishna River basin, and the criteria include six parameters, namely, slope, elevation, flow accumulation, land use and land cover, rainfall intensity, and distance from the drainage Table 1 Sources of data collection of different process Factor

Source

Download URL

Elevation (ASTER DEM)

USGS earth explorer

https://earthexplorer.usgs.gov/

Land use and landcover (Landsat 8)

USGS earth explorer

https://earthexplorer.usgs.gov/

Rainfall intensity

IMD Pune

https://www.imdpune.gov.in/

Soil (global soil maps)

FAO website

https://www.fao.org/home/en

Slope, flow accumulation, distance from drainage network

Generated using DEM in ArcMap 10.7.1 (licensed version)

–

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

267

(i) Fig. 2 a Digital elevation model, b slope map, c flow direction map, d flow accumulation map, e land use and land cover map, f rainfall intensity map, g stream order map, h distance from drainage map, i soil map

268

S. Kumar et al.

Table 2 Relation between factors that influence flood risk Factor changing

Major effect

Minor effect

Factor rate (FR)

Flow accumulation (F)

LULC

S

1.5

Slope (S)

LULC, F

–

2

Land use and landcover (LULC)

F, R

S

2.5

Rainfall intensity (R)

F

LULC

1.5

Elevation (E)

LULC, F, R

S

3.5

Distance from drainage network (D)

LULC, F, R

S

3.5

Soil (A)

LULC, F, S

–

3

network. Secondly, the alternatives are defined. In this study, the alternatives refer to the hazardous areas within the study area. Next, priority is established amongst the criteria and alternatives using pairwise comparison. The pairwise comparison involves comparing each criterion or alternative with every other criterion or alternative and assigning a score that reflects the relative importance or preference between them. Later, the consistency amongst the pairwise comparison is checked to ensure the validity of the comparisons. Next, the relative weights are evaluated from the pairwise comparisons, and the overall priorities for the alternatives are calculated based on these weights. By employing the methodology presented by Kourgialas and Karatzas (2016), the weightage of each parameter is determined. The last step involves conducting a sensitivity analysis to evaluate how variations in the weightage allocated to each parameter affect the robustness of the outcomes. Overall, the AHP method provides a step-by-step approach to decision-making, and it can be applied to various complex problems, including flood hazard mapping. The method is effective in identifying the most important factors that influence the decision and assigning appropriate weights to these factors, thereby providing a systematic and transparent decision-making process. In this study, a GIS-based weighted linear combination approach was used to combine the seven maps that were generated as shown in Eq. 1. The technique involved assigning a percentage weight to each factor and multiplying it by the corresponding factor. The resulting products were summed up to create the final map of the hazardous area. Figure 3 illustrates the detailed methodology that was employed in this process. Flood Risk Map =



X i Wi = F Wr + SWs + L Wl

+ RWr + E We + DWd + AWa

(1)

where X i = the map of each parameter i, W i = the weight of each parameter i, F = flow accumulation factor, S = slope factor, L = land use factor, R = rainfall intensity factor, E = elevation factor, D = distance from drainage network. A = slope map and the respective weights is shown in Table 3 and total flood hazard area is shown in Table 4.

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques

269

Fig. 3 Flowchart of methodology Table 3 Weightage of factors causing flood risk Factors

Range of effect

Description

Weight of effect (A)

Rate (B)

Weighted rating (A × B)

Total weight

Elevation

153–332

Very high

5

1.5

7.5

22.5

Slope

Flow accumulation

86–152

High

4

6

47–85

Moderate

3

4.5

27–46

Low

2

3

3–26

Very low

1

1.5

22–57

Very high

5

2

2

12–21

High

4

4

7–11

Moderate

3

6

4–6

Low

2

8

0–3

Very low

1

10

601,227–1,110,961

Very high

5

1.5

7.5

304,971–601,226

High

4

6

161,199–304,970

Moderate

3

4.5

52,282–161,198

Low

2

3

0–52,281

Very low

1

1.5

30

22.5

Percentage (%)

8.62

11.49

8.62

(continued)

270

S. Kumar et al.

Table 3 (continued) Factors

Range of effect

Landuse and landcover

Water Body Vegetation Built-up area

Low

Rainfall intensity

Weight of effect (A)

Rate (B)

Very high

5

3

Moderate

3

9

2

6

Barren land

Very low

1

1134–1202

Very high

5

1075–1133

High

4

Weighted rating (A × B)

Total weight

Percentage (%)

15

33

12.64

3 1.5

7.5

1026–1074

Moderate

3

4.5

Low

2

3

876–977

Very low

1

Very high

5

High

4

22.5

Moderate

3

13.5

2

9

Very low

1

4.5

Loam

Moderate

3

Low

2

6

Clay

Very low

1

3

High

3

9

Area (in Km2 ) 4.603

Percentage of area (%)

299.9

29.63

616.206

60.87

91.312

9.02

1012.24

6.9

0.4

Low Total

25.86

18

Moderate Very low

67.5

18

Low

Flood impact

8.62

1.5 4.5

Sandy clay loam

Table 4 Flood hazard area

22.5

6

978–1025 Distance from drainage

Soil

Description

100

4 Results and Discussion The generation of results in this study involved the preparation of seven input maps: Digital Elevation Model (DEM), Slope Map, Flow Accumulation Map, Land Use and Land Cover Map, Rainfall Intensity Map, Soil Map, and Distance from Drainage Map as shown from Figs. 4, 5, 6, 7,8, 9 and 10. To ensure an equal class distribution for all maps and account for their relative importance, each input map was reclassified into five classes using the reclassify tool in ArcGIS software. This allowed us to spread the weightage values and obtain a more accurate output of the final hazard map shown in Fig. 11. The reclassified maps are listed below. The reclassified maps are as follows.

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques

Fig. 4 Reclassified DEM map

Fig. 5 Reclassified slope map

271

272

Fig. 6 Reclassified LULC map

Fig. 7 Reclassified rainfall intensity map

S. Kumar et al.

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques

Fig. 8 Reclassified distance from drainage map

Fig. 9 Reclassified soil map

273

274

Fig. 10 Reclassified flow accumulation map

Fig. 11 Final flood hazard map

S. Kumar et al.

Flood Hazard Mapping for Amaravati Region Using Geospatial Techniques

275

After the reclassification of these maps into the required number of classes (say 5 here), all these maps are given as input to the Weighted Overlay Table tool in ArcGIS software with the respective weights of each map. This tool makes all these maps overlay on each other with respect to the weightages given (sum of all the given weightages should be 100) and give the output of the final hazard map based on all the given inputs, which are a part of AHP. Finally, the final hazard map is shown in Fig. 11. The study aimed to assess flood risk in a selected study area of 1012.24 Km2 and identify areas vulnerable to flooding. Our modeling results revealed a high-hazard flood zone covering 4.603 Km2 , indicating the significant vulnerability of this region. Conversely, a vast area of 616.206 Km2 was observed to be less prone to flooding, reflecting the impact of human interventions, such as infrastructure development and flood management measures. These findings have significant implications for flood risk reduction and mitigation strategies in the study area. The identification of high-risk flood zones can inform targeted investments in flood protection measures, reducing the potential impact of future flood events on people’s lives and livelihoods. Furthermore, our study highlights the need for continued investment in flood risk reduction measures to minimize the impact of natural disasters on vulnerable communities. Overall, our results provide valuable insights into flood risk assessment and management and can inform future policy decisions aimed at improving the resilience of communities in flood-prone regions. However, there are lack of monitoring datasets for calibration and validation is one of the limitations of the present study; However, demonstrates the power of geospatial technology in flood risk analysis and decision-making. The comprehensive flood hazard and flood shelters model developed in ArcGIS can be easily used by novice GIS users, allowing for effective flood management and disaster response. With the increase in rainfall due to climate changes leading to more floods, the use of geospatial data and AHP process will provide a valuable source of analysis.

5 Conclusions Floods are natural phenomena that cannot be prevented, but human activity is exacerbating their impact. This study aimed to create a flood hazard map of the Amaravati region using remote sensing and GIS techniques. By analyzing data from seven different factors, we identified low-lying areas near the stream network as particularly susceptible to flooding during heavy rains. However, our focus area, Amaravati, experiences less rainfall and is less vulnerable to floods. This study demonstrates the power of geospatial technology in flood risk analysis and decision-making. The comprehensive flood hazard and flood shelters model developed in ArcGIS can be easily used by novice GIS users, allowing for effective flood management and disaster response. The final flood hazard map generated from our analysis can inform preventive measures and evacuation plans in flood-prone areas while building designs can be modified to reduce human and financial losses. The methodology adopted in the

276

S. Kumar et al.

present study using the AHP approach can be widely applied to generate flood hazard maps for other regions. This study supports the conclusion that remote sensing and GIS technology can provide valuable information to operational users for planning flood-related emergency responses. Ultimately, this study highlights the importance of utilizing technology to improve disaster preparedness and response to natural disasters.

References 1. Mudashiru RB, Sabtu N, Abustan I, Balogun W (2021) Flood hazard mapping methods: a review. J Hydrol 603:126846. https://doi.org/10.1016/j.jhydrol.2021.126846 2. Alarifi SS, Abdelkareem M, Abdalla F, Alotaibi M (2022) Flash flood hazard mapping using remote sensing and GIS techniques in Southwestern Saudi Arabia. Sustainability 14(21). https:// doi.org/10.3390/su142114145 3. Elkhrachy I (2015) Flash flood hazard mapping using satellite images and GIS tools: a case study of Najran City, Kingdom of Saudi Arabia (KSA). Egypt J Remote Sens Space Sci 18(2):261–278. https://doi.org/10.1016/j.ejrs.2015.06.007 4. Feng Q, Gong J, Liu J, Li Y (2015) Flood mapping based on multiple endmember spectral mixture analysis and random forest classifier—the case of Yuyao, China. Remote Sens 7(9). https://doi.org/10.3390/rs70912539 5. Surwase T, Manjusree P, Nagamani PV, Jaisankar G (2019) Novel technique for developing flood hazard map by using AHP: a study on part of Mahanadi River in Odisha. SN Appl. Sci. 1(10):1196. https://doi.org/10.1007/s42452-019-1233-6 6. Sharma VS, Naithani BP, Singh M (2014) Remote sensing and GIS approach for hazard vulnerability assessment of upper Alaknanda Basin, Garhwal Himalaya (Uttarakhand), India. In: Landscape ecology and water management. Tokyo, pp 273–286. https://doi.org/10.1007/9784-431-54871-3_20 7. Morales FJ, De Vries W (2021) Establishment of natural hazards mapping criteria using analytic hierarchy process (AHP). Front Sustain 2:667105. https://doi.org/10.3389/frsus.2021.667105 8. Ghosh A, Kar SK (2018) Application of analytical hierarchy process (AHP) for flood risk assessment: a case study in Malda district of West Bengal, India. Nat Hazards 94(1):349–368. https://doi.org/10.1007/s11069-018-3392-y 9. Mahapatra M, Ramakrishnan R, Rajawat AS (2015) Coastal vulnerability assessment using analytical hierarchical process for South Gujarat coast, India. Nat Hazards 76(1):139–159. https://doi.org/10.1007/s11069-014-1491-y 10. Gupta L, Dixit J (2022) A GIS-based flood risk mapping of Assam, India, using the MCDAAHP approach at the regional and administrative level. Geocarto Int 37(26):11867–11899. https://doi.org/10.1080/10106049.2022.2060329

GIS and RS-Based Soil Erosion and Sediment Yield Modelling in Manikpur, Chhattisgarh, India B. Himajwala and A. D. Prasad

Abstract The Manikpur coalfield is 300 km2 and is in the Korba district of the Indian state of Chhattisgarh. In India, one of the most severe issues is soil erosion. Calculating exact soil erosion over a specific time is extremely difficult. Huge quantities of mining waste are typically released as over burden dump (OBD) materials by opencast mines. They impact agriculture since they are prone to soil erosion, sedimentation, and poor water quality. An empirical equation of the Revised Universal Soil Loss Equation (RUSLE) and the Sediment Delivery Distributed Model (SEDD) were employed to quantify soil erosion and sediment yield. The findings of these equations were contrasted to those of direct field measurements obtained around the opencast region utilizing an appropriate suspended sediment sampler. The maximum soil erosion value obtained is 79.2 tons/ha/year, and the maximum sediment yield is 57.92 tons/ha/year, according to the findings of this study. The sediment load values from these two models follow the same trend. These models were evaluated for performance and found to be effective, with values of simulated and observed sediment yields R2 = 0.81 and RRMSE = 0.66. The observations revealed that most of the area has a low slope gradient. The opencast mine and overburden dump area have a slight erosion potential due to the higher slope inclination. According to the study, GIS is an effective tool for modelling soil erosion potential and sediment output. Keywords GIS · OBD · SEDD

1 Introduction Soil erosion is “the removal of soil from the exposed surface of land due to the impact of raindrops and running water, including rainfall-runoff, melting snow and ice, wind or gravity, and human activities.” Soil erosion is mainly caused by deforestation, cattle overgrazing, building, and mining operations. The soil sheet is removed during opencast mining, and over burden dumps (OBD) are built up from the fractured B. Himajwala (B) · A. D. Prasad Department of Civil Engineering, NIT Raipur, Raipur, CG 492010, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_20

277

278

B. Himajwala and A. D. Prasad

rock [1]. The dumping materials are dispersed on the ground as OBD; they cover a significant amount of land, preventing it from serving its intended purpose and, as a result, degrading soil quality [2]. On the other hand, the mining sector has a bad reputation since it can endanger public health and safety and harm a local, regional, and global scale; the environment includes land, soil, water, and trees [3]. They cover a vast region of ground that is no longer used for its intended purpose and has poorer soil quality [4]. Surface mining areas can produce 100–2,000 times more sediment than forested areas and 10 times more sediment than grazing areas [5]. Physical and human factors influence sediment generation and delivery [6]. RUSLE utilizes a combination of Geospatial Information System (GIS) and Remote Sensing (RS) to more precisely estimate annual soil loss in the catchment (GIS).

2 Study Area Manikpur opencast mine is situated in the Korba Coalfield, in the Korba district of Chhattisgarh, India. The study area depicted in (Fig. 1) is 115.4 km2 area. The region’s terrain is plain, with open mines, forests, and streams, among other things. The average yearly temperature in the area is between 45 and 55 °C.

a

b

Fig. 1 Study area with sample collection locations and soil erosion observed in the studied region: a Rill and inter-rill erosion at location no.17, b stream bank erosion at location no. 3

GIS and RS-Based Soil Erosion and Sediment Yield Modelling …

279

2.1 Data Used From 2011 to 2021, daily rainfall data were obtained from (SECL, Manikpur) for six rain gauge stations in the study area: Manikpur, Aurai, Nagoi, Samarkana, and Pondi. The velocity of a stream was determined using current meter revolutions at various cross-sections using surface water samples taken in the Manikpur coalfield region at 13 locations. Digital Elevation Model (DEM) data from the Shuttler Radar Topographic Mission produced a model with a cell size resolution of 30 m × 30 m (SRTM, USGS). The data on land use, land cover, and soil texture were provided by the Chhattisgarh Council of Science and Technology (CCOST), Raipur, Chhattisgarh.

3 Methodology This describes how to use the direct approach as well as the RUSLE model to calculate the yearly average soil loss rate and sediment yield. The graphs illustrate sediment production and soil degradation for the chosen parameters in (Fig. 2).

3.1 Soil Erosion and Sediment Yield Modelling Using Empirical Equations In this study, the RUSLE model is used to determine how much soil is lost to the watershed. Climate, soil qualities, topography, and land cover characteristics are the primary elements impacting soil erosion, and the RUSLE model established an equation to reflect these factors [7]. The RUSLE is an empirical soil loss forecasting

DIRECT METHOD

RUSLE

RUSLE

LULC

SEDIMENT CONCENTRATION

SEDIMENT LOAD

RESULTS COMPARISION & ANALYSIS

ARUSLE = Average Annual Soil loss SDR = Sediment Delivery Ratio

Fig. 2 Methodology

C FACTOR

DEM

P FACTOR

SOIL TEXTURE

LS FACTOR

SEDIMENT YIELD = ARUSLE

K FACTOR

SDR

RAINFALL

R FACTOR

280

B. Himajwala and A. D. Prasad

model which predicts mean yearly soil loss and sediment yield [8]. quantitatively forecasts soil erosion. It evaluates the yearly averages of inter-rill and rill erosionrelated soil loss and sediment [9]. This is how the formula is written: A = R × K × LS × C × P

(1)

R is the rainfall-runoff erosivity factor (MJ mm/ha*h*y); K is the soil erodibility factor) (t* ha* h/ha*MJ*mm); ARUSLE is estimated average soil erosion (t/ha/year); LS is an expression for slope length and gradient, whereas crop management is denoted by the letter C. P is a consideration of support practice (0–1). Rainfall Erosivity Factor (R): The R factor is an MJ.mm/(ha.hr.year) erosion index [10]. The R factor estimates the volume and rate of flow connected to events causing rainfall [11]. The KE of a storm in the metric unit is calculated by equation developed [7].   mm ∗h∗y R = 79 + 0.363 × P MJ ha

(2)

Soil Erodibility Factor(K): K Factor is controlled by organic material, soil conditions, and soil structure all affect the K factor. The soil’s percolation ability and high stability of the soil impact the K factor’s value. The soil’s organic matter content and soil distribution of particle sizes are the primary soil variables needed to compute the soil erodibility K-value (sand, silt, and clay) [9]. The main types of soil in the research area are Loamy, clayey, and loamy skeletal. The K variable values are obtained from literature [12], and the geographic area under each soil class has been given. Slope Length and Steepness Factor (LS): RUSLE studies an advanced method for finding the slope length factor. DEM’s equation for calculating the LS factor is as follows [13]: L = (λ/22.1)0.5

(3)

S = 10 × sin θ + 0.03 if slope < 9%

(4)

S = 16.8 × sin θ − 0.05 if slope > 9%

(5)

Cover Management Factor(C): The cover management component considering the most practicably achievable calculated yearly soil loss of the catchment came within a range varied from classes in the study region in comparison to earlier studies, C is arguably the far more significant of the RUSLE variables [14]. Other land use categories for allocating C factor values are listed in Table 1. Support Practice Factor (P): These methods have a negative effect on soil loss by managing the runoff’s flow pattern and decreasing its volume and velocity [15]. According to previous research [16], the values 0.28 and 1 were attributed to paddy and non-paddy fields, respectively, on the P factor map.

GIS and RS-Based Soil Erosion and Sediment Yield Modelling … Table 1 LULC of the studied region with C values

281

LULC

C

Area (ha)

% Area

Cropland

0.28

8384.77

36.7

Built up

0.23

969.63

4.5

Forest

0.13

8040.62

36.4

Water bodies

0

311.4

1.4

Wastelands

0.14

270.26

20.3

Sediment Delivery Ratio by GIS-Based Sediment Delivery Distributed Model. It is the proportion between a region’s sediment production and overall erosion. The ratio of sediments distribution for the catchment, SDR, is approximated using the following equation [17] as a function of travel time. SDRi = e(−β × ti) ti =

k i=1

(λn/ V i)

(6) (7)

The stream velocity is determined from Manning’s equation, which is a combination of inclination of the ground surface and characteristics of the forest cover [18]. Vi = ai × (Si )0.5

(8)

where: it is the travel time (h), β is a parameter particular to a watershed. λn = length of a flow route segment n, m, ai = surface roughness coefficient for cell n, m/s, S i = slope for cell n, m/m.

3.2 Direct Method for the Estimation of Sediment Yield In direct method analysis, surface water samples were taken with a bottle-type sampler at thirteen locations across the Manikpur coalfield area to determine the suspended sediment yield. The stream discharge is measured using the area velocity approach. A filtration test is performed on each sample to measure the suspended sediment concentration. Sediment load = Discharge(Q) × Suspended sediment concentration(C)

(9)

282

B. Himajwala and A. D. Prasad

4 Results and Discussion The erosion map is obtained by multiplying the raster images utilizing the ArcGIS raster converter. The calculated yearly soil degradation of the catchment fell within a span of 0–79.2 tons per hectare per year, as shown in (Fig. 7) having an average value of 0.30. The soil loss map is categorized in accordance with the erosion risk groups proposed by [16] for Indian circumstances, as described in Table 2. Rainfall-runoff erosivity of the catchment was found high at 1 and 3 point around Korokoma shown in Fig. 2 and low at point 16 around Bhilaikurd. In the study area, Loamy soil is present at 1, 3 points around Korokoma and 8, 12, 13, 16 points around Manikpur, Nakhtikar, Kudri, and Bhilaikhurd, respectively, which is undergone moderate to severe soil erosion. K factor map as shown in Fig. 4 with an average value of 0. Steep and rough slopes encompass the opencast mine and overburden dumps (Fig. 5) . Cover management is dimensionless, as demonstrated in Fig. 6. Around 73.1% of the studied area was comprised of vegetative covers, such as woods and croplands. In comparison, only roughly 26.9% comprised built-up regions, water bodies, and wastelands. The study area’s P-factor image, as displayed in Fig. 7, was created by assigning Pvalues to various land use classifications. The sediment arrival rate shown in Fig. 9 has a high spatial distribution at the dam site and outlet, especially in areas with high slopes. This was due to a nearby outlet time of concentration being low and flow velocity being high at sloppy locations. Slight spatial variability in sediment yield throughout the whole catchment region (Fig. 8) ranges from 0 to 57.93 t/ha/year. Due to enhanced flow accumulation downslope from the source to the catchment outlet, high sediment production occurs in river valleys and tends to grow downstream (Fig. 10). Table 2 Distribution of various erosion classes

Soil erosion class

Area (ha)

i), the resulting statistics for the MK-test are as follows.

292

K. N. Loukika et al.

Fig. 2 IMD grid points and groundwater observation wells

Groundwater Level data (1996-2018)

Seasonal groundwater level analysis (Monsoon, Post-monsoon Kharif , Post-monsoon Rabi, Pre-monsoon)

Trend analysis Spatial groundwater level maps using interpolation Mann Kendall Test

Results Fig. 3 Methodology flowchart

Groundwater Level Trends Over Southern India

S=

n n−1  

293

sgn(x j − xi )

(1)

i=1 j=i+1

With   ⎧ ⎫ ⎨ 1, i f x j − xi  > 0 ⎬ sgn(x j − xi ) = 0, i f x j − xi = 0 . ⎩ ⎭ −1, i f x j − xi < 0

(2)

If the x i values are independent and randomly ordered and n > 10, the statistic S follows a normal distribution with a mean of zero and a variance given by the following equation.  V ar (S) = n(n − 1)(2n + 5) −

m 

ti i (i − 1)(2i + 5) /18

(3)

i=1

In this context, m represents the number of tied groups in the entire data set, while t i represents the total number of data points in the ith tied group. The equation used to calculate the standardized test statistic Z MK is as follows:

ZMK

⎧ S−1 ⎪ ⎨ √V ar (S) f or S > 0 = 0 f or S = 0 ⎪ ⎩ √ S+1 f or S < 0 V ar (S)

(4)

A two-tailed test is used to determine if the null hypothesis should be rejected for a specific significance level α (90, 95), where Z MK must be greater than Z 1−α/2 . .The trend will be identified as increasing or decreasing based on the sign of the test statistic. The magnitude of the trend is determined using the Theil-Sen approach (TSA) [9].  β = median

 x j − xi for allj > i. j −i

(5)

3.2 Results Figure 4 shows the spatial variation of mean monthly seasonal groundwater levels for the period 1996–2018. The groundwater level is high in the pre-monsoon season compared to other three seasons. During the monsoon period from 1996 to 2018, Coimbatore in Tamil Nadu had a mean monthly groundwater level of 12.17 m. In the

294

K. N. Loukika et al.

Post-Monsoon Rabi period, the mean monthly groundwater level was 10.7 m, while in the Post-Monsoon Kharif period it was 11.36 m. The Pre-monsoon mean monthly groundwater level was 12.4 m, which was higher compared to other observation wells. Depth to below groundwater level in Post-Monsoon Kharif is less in Jajapur which is in Odisa state. Trends in groundwater level for the Monsoon, Post-monsoon Kharif, Postmonsoon Rabi and Pre-monsoon seasons were analyzed using the MK test is shown in the Fig. 5. The study utilized data on piezometric levels from 181 observation wells during the period from 1996 to 2018. Monsoon season In the Telangana state, Nalgonda has exhibited a decreasing trend, while all other observation wells showed no trend in the Monsoon season. Similarly, in Andhra Pradesh, five wells located in Vizianagaram, East Godavari, Srikakulam, Nellore, and Visakhapatnam showed a decreasing trend, while the remaining wells showed no trend during the Monsoon season. In Chattisgarh state, Bastar, Dantewada, Raigarh, Uttar Bastar Kanker, Narayanpur, and Raipur regions exhibited a decreasing trend, while other wells showed no trend. In Karnataka, Bijapur, Davanagere, Dharwad,

Fig. 4 Spatial variation of Groundwater levels for the period 1996–2018

Groundwater Level Trends Over Southern India

295

Fig. 5 The spatial distribution of Z-values and trends of grid points in groundwater levels across southern India during the period 1996–2018

Gulbarga, Haveri, Raichur, Shimoga exhibited a decreasing trend, while Chitradurga showed an increasing trend. From Kerala state, Wayanad showed a decreasing trend. Maharashtra state had a decreasing trend in the Aurangabad, Chandrapur, Jalna, Kolhapur, and Wardha regions. Similarly, in Tamil Nadu, 13 wells exhibited a decreasing trend. Additionally, the decreasing trend was observed in Cut tack, Dhenkanal, Gajapati, Ganjam, Kandhamal, Kendrapara, Koraput, and Malkangiri from Odisha state. Finally, Thiruvananthapuram in Kerala exhibited a decreasing trend. Post-monsoon Rabi Observations from several states in India indicate that some wells are showing a decreasing trend. For instance, Andhra Pradesh had four such wells, Chattisgarh and Kerala had three and five wells, respectively, while Karnataka had 13 wells with a decreasing trend. Maharashtra had one well in Greater Bombay that exhibited an increasing trend, but eight other wells had a decreasing trend. Odisha had the highest number of wells with 21 showing a decreasing trend, while Tamil Nadu had five

296

K. N. Loukika et al.

wells with the same pattern. Only one well from Telangana (Khammam) displayed a decreasing trend. Post-monsoon kharif In Andhra Pradesh, four wells located in Vizianagram, East Godavari, Srikakulam, and Visakhapatnam have shown a decreasing trend. Similarly, the only well located in Chattisgarh, Dantewada, has also exhibited a de creasing trend. In Karnataka, five wells and two wells in Kerala, namely Thiruvarur and Kasaragod respectively, have also shown a decreasing trend pattern. Additionally, four wells in Maharashtra, five wells in Odisha, and six wells in Tamil Nadu have exhibited decreasing trend patterns. However, in Telangana, Mahabubnagar has shown an increasing trend pattern. Pre-monsoon There has been a decreasing trend observed in several districts of Andhra Pradesh, including Vizianagaram, East Godavari, West Godavari, Srikakulam, and Visakhapatnam. Bijapur in Chattisgarh has experienced an increasing trend, while Dantewada, Bastar, Narayanpur, and Uttar Bastar Kanker have shown a decreasing trend. Only one well located in Bijapur, Karnataka, has exhibited a decreasing trend. Similarly, 20 wells in Odisha and 4 wells in Kerala have exhibited a decreasing trend pattern. The trend has been increasing in Greater Bombay and Buldana in Maharashtra, while Kolhapur, Garhchiroli, and Gondiya have exhibited decreasing trend patterns. Additionally, 20 wells in Odisha and 7 wells in Tamil Nadu have shown a decreasing trend. Mahabubnagar and Hyderabad have shown an increasing trend pattern, while Khammam has experienced a decreasing trend.

4 Conclusions The study found that there was a notable decrease in water levels in several observation wells across India during different seasons. Specifically, 47 wells experienced a declining trend during the monsoon season, 61 wells during the Post-Monsoon Rabi season, 27 wells during the Post-Monsoon Kharif season, and 45 wells during the Pre-Monsoon season. Only one well showed an increasing trend across the eight states studied. The results showed that there were more declining trends during the Post-Monsoon Rabi and Pre-Monsoon seasons. The insights from this present study were useful for the policy makers and stakeholders especially in the irrigation and agricultural sector in an inevitable manner for getting maximum benefits.

Groundwater Level Trends Over Southern India

297

References 1. Bhanja SN, Mukherjee A, Rodell M, Wada Y, Chattopadhyay S, Velicogna I, Panglauru K, Famiglietti JS (2017) Groundwater rejuvenation in parts of India influenced by water-policy change implementation. Sci Rep 7(1). https://doi.org/10.1038/s41598-017-07058-2 2. Miro EM, Famiglietti JS (2018) Downscaling GRACE remote sensing datasets to high— resolution groundwater storage change maps of California’s Central Valley. Remote Sensing 10(1) 3. Patle GT, Singh DK, Sarangi A, Rai A, Khanna M, Sahoo RN (2015) Time series analysis of groundwater levels and projection of future trend. J Geol Soc India 85:232–242. https://doi.org/ 10.1007/s12594-015-0209-4 4. Kendall MG (1955) Rank correlation methods. Charles Griffin, London 5. Patle GT, Singh DK, Sarangi A, Rai A, Khanna M, Sahoo RN (2013) Temporal variability of climatic parameters and potential evapotranspiration. Indian J Agric Sci 83(4):518–524 6. Jagadeesh P, Anupama C (2013) Statistical and trend analyses of rainfall: A case study of Bharathapuzha river basin, Kerala, India. ISH J. Hydraulic Eng. 20(2):119–132. https://doi.org/ 10.1080/09715010.2013.843280 7. Mann HB (1945) Non-parametric tests against trend. Econ-metrica 13:245–259 8. Kendall MG (1962) Rank correlation methods. HafnerPublishing Company, New York 9. Kumar KS, Rathnam EV (2009) Analysis and prediction of groundwater level trends using four variations of Mann Kendall tests and ARIMA modelling. J Geol Soc India 94:281–289. https:// doi.org/10.1007/s12594-019-1308-4

Impact of Climate Change on Streamflow Over Nagavali Basin, India Nageswara Reddy Nagireddy and Keesara Venkatareddy

Abstract Natural disasters such as cyclones pose a serious threat to the Indian coast. The Nagavali basin is an interstate east-flowing river that supports agricultural and domestic water demands in Koraput, Kalahandi, and Rayagada districts in Odisha, as well as Vizianagaram and Srikakulam in Andhra Pradesh. Furthermore, this basin is especially prone to cyclones generated by low-pressure depressions in the Bay of Bengal. The present research aims to simulate the streamflow of the Nagavali basin using the SWAT model and to provide historical information as well as future changes within streamflow in response to climate change. Calibration (1991–2005) and validation (2006–2014) of the SWAT model showed a satisfactory for monthly streamflow. The downscaled bias-corrected geographically separated the NASA NEX-GDDP dataset was employed for simulating the future streamflow under two RCP scenarios, 4.5 and 8.5. The analysis periods were divided into 27-year blocks that included a historical period (1980–2005) as well as three future periods, namely near term (2022–2047), the middle future (2048–2073), and the far future (2074–2099). Under both scenarios, the CNRM-CM5 model predicted a rising trend in precipitation and streamflow. Under RCP 4.5, the CNRM-CM5 model showed the greatest percentage change of 22.06% in the far future, with a corresponding streamflow change of 36.26%. Under RCP 8.5, the BNU-ESM, CNRM-CM5, and IPSLCM5A-MR models represent percentage change in precipitation ranging from 22.09 to 26.28% and corresponding streamflow ranging from 33.81 to 45.11%, respectively, in the far future. Peak discharges are estimated at various return periods with Log Pearson Type III probability distribution. The estimated peak discharge (5626 m3 /s) of the 100-year return period was observed in 2006. This study findings are useful to water resource managers. Keywords Nagavali basin · Climate change · Flood frequency analysis · SWAT

N. R. Nagireddy (B) · K. Venkatareddy Department of Civil Engineering, NIT Warangal, Hanamkonda, Telangana, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_22

299

300

N. R. Nagireddy and K. Venkatareddy

1 Introduction Climatic situations over the twenty first century forecasted by Global Climate Models (GCMs) revealed a warming trend around the world. However, the severity and nature of these effects vary greatly across regions. Floods and droughts will become more common in the majority of the globe, resulting in lower crop productivity and an increase in the risk of hunger and poverty [1]. The consequences of warming temperatures on hydrological processes and the management of water resources have become a major cause of attention in recent years. Numerous simulations, studies, and experiments based on general circulation models (GCMs) predicted an increase in temperature as greenhouse gas levels increased. Climate processes are expected to worsen as a result, with serious effects for agricultural and energy output, ecosystems, flooding, and drought. For long-term strategic management of a nation’s water assets in the context of evolving changing climate consequences, it is essential that the consequences of climate change be assessed using high temporal and spatial resolution across the basin scale. Wagner et al. [2] employed SWAT to examine the consequences of climate and land use changes in the Western Ghats. India is an important economically developing nation, with almost sixty percent of its population relying directly on climate-sensitive sectors such as farming and fisheries [3]. The spatial patterns of Indian rainfall suggested that the west coast and northeast India would receive the most rain under both the A2 and B2 IPCC scenarios [4]. There will be plenty of water in the Godavari, Brahmani, and Mahanadi basins, but catastrophic flooding is possible [5]. The Indian summer monsoon rainfall, which occurs from June to September accounting for more than 75% of total precipitation in Indian river basins [6]. So, the accurate projections of the summer monsoon is important, because it is the most important drivers for climate change policies and adaptation for Indian river basins [7]. Raghavan et al. [8] investigated the NEX-GDDP dataset, which includes statistically downscaled CMIP5 past and future climate forecasts at daily and 25 km spatial resolution. They concluded, surface temperatures in Southeast Asia will rise by greater than 3.5 °C in the future. Rao et al. [9] forecasted future shifts in extreme rainfall over south India with high-resolution NEX-GDDP datasets. This NEX-GDDP dataset replicates mean rainfall fairly well. Rainfall may increase up to 11% in the near and up to 38% in the far future, under RCP 4.5 and 8.5 scenarios. According to future climate predictions, the severity, as well as the timing of rainfall extremes in most parts of south India, may increase throughout the monsoon season under both scenarios. Nagireddy et al. [10] evaluated the climate change impact on streamflow and sediment in the Nagavali and Vamsadhara watersheds using CMIP6 climate models data and SWAT model. They found the range of projected changes in the percentage of average annual precipitation and average temperature ranged from 0 to 41.7% and 0.7 to 2.7 °C, respectively, indicating a warmer and wetter climate in the Nagavali and Vamsadhara watersheds. The occurrence and severity of peak discharges connected with individual events are used to give significant details on the probability of floods. The flood frequency assessment depends on existing data from each year’s independent flood discharge

Impact of Climate Change on Streamflow Over Nagavali Basin, India

301

maxima. Dams, bridges, and flood prevention structures can all benefit from understanding frequency and magnitude relationships. Griffis and Stendinger [11] estimated maximum yearly precipitation and discharges using the Log Pearson III distribution. In the case of Indian rivers, the Log Pearson-III distribution produces reliable outcomes for calculating peak discharges up to a 100-year return interval [12, 13]. The primary goal of this research is to evaluate the effects of climate change on streamflow in the Nagavali basin using the SWAT model and to calculate the peak discharges at distinct return periods using Log Pearson type III probability distribution.

2 Materials and Methods 2.1 Study Area The Nagavali basin is an intrastate east-flowing river basin residing adjacent to the Godavari and Mahanadi basins (Fig. 1). The Nagavali river originate in the vicinity of Lakhbahal village in Odisha’s Kalahandi district. It flows 256 km before joining the Bay of Bengal. The drainage area of the Nagavali basin is 9200 km2 .

Fig. 1 Study area map

302

N. R. Nagireddy and K. Venkatareddy

2.2 Datasets Spatial datasets DEM, LULC and Soil map used in the present study found in [14]. NEX-GDDP dataset obtained from Coupled Model Inter-comparison Project Phase 5 (CMIP5) simulations under two RCP emission scenarios, 4.5 and 8.5 were used. It has a spatial resolution of 0.25°. These datasets include a set of global, high-resolution, bias-corrected climate change forecasts which may be employed for evaluating the consequences of climate change at finer scales.

2.3 Model Setup and Methodology The SWAT model is a broad, ongoing system constructed by the U.S. Department of Agriculture to analyse and predict hydrologic processes over long time periods. It is intended to simulate daily surface runoff, sediment yield, nutrient and chemical loads in farming operations. The model has been exhaustively described by [15]. To begin, the SWAT model was created by projecting a DEM, LULC map, and soil map into the WGS 1984 UTM 44N projection. Using the QSWAT tool on the QGIS interface, the Nagavali basin was split into 34 subbasins and 2153 hydrological response units (HRUs) depending on components like soil uniformity, landuse, slope, and the threshold area of 100 ha. The SWAT model was initially calibrated and validated with streamflow. To evaluate the model performance, the objective function Nash Sutcliffe efficiency (NSE) [16] was used. It was determined that the model performance was satisfactory if the NSE was greater than 0.5 [17]. Based on previous reported research, we have selected three CMIP5 models in this study such as CNRMCM5, BNU-ESM and IPSL-CM5A-MR models.

3 Results and Discussions 3.1 Flood Frequency Analysis Annual Maximum Series (AMS) data from the Srikakulam station over a 29-year period (1990–2018) were used for the analysis of flood magnitude and frequency in the Nagavali basin. The observed yearly Maximum Series in Table 1 were used to calculate a designed flood. Peak flows for 2-year, 5-year, 10-year, 25-year, 50-year, and 100-year return periods were calculated employing the Log Pearson type III probability distribution, and the calculated discharges are presented in Table 2. The highest (5626 m3 /s) calculated discharge of 100 year return period was observed in the year of 2006 over Nagavali basin, the flood frequency analysis for 29 years of annual maximum data showed the 13 events fall within the 2-year return period,

Impact of Climate Change on Streamflow Over Nagavali Basin, India

303

12 events fall under the 5-year return period, 2 events fall under the 10-year return period, and one event falls under the 50- and 100-year return periods. Table 1 Annual maximum series of streamflow and their rank Year of occurrence of event

Ranked maximum streamflow (m3 /s)

Rank

2006

5624.739

1

2014

4224.446

2

2013

2141.784

3

1992

2012.6

4

1995

1917

5

2012

1900

6

2008

1703.286

7

1994

1640.2

8

1991

1497

9

2011

1442.89

10

2018

1389.468

11

2009

1374.639

12

2010

1338.666

13

1996

1284

14

2015

1200

15

2017

1191

16

2003

1086.72

17

2007

1014.324

18

1990

1009.5

19

2000

852.94

20

2005

796.293

21

2016

772.102

22

2001

758.841

23

2004

658.59

24

1997

577.8

25

1998

451.05

26

1999

370.34

27

1993

337.52

28

2002

301.176

29

304

N. R. Nagireddy and K. Venkatareddy

Table 2 Estimated streamflow using log pearson type III distribution Return period

Skew coefficient K(0.0805)

Streamflow Q(m3 /s)

2-year

−0.014

1114

5-year

0.837

1979

10-year

1.290

2686

25-year

1.778

3735

50-year

2.097

4630

100-year

2.386

5627

3.2 Calibration and Validation The observed streamflow data at the Srikakulam station was used to calibrate the SWAT model over the Nagavali basin. It should be emphasized that for a complicated model like SWAT, where there are many parameters to estimate, manually achieving the ideal parameter set is challenging. As a result, auto calibration with parameter optimization was performed to obtain the best SWAT model parameters. 15 parameters influencing the streamflow were considered for streamflow calibration. Table 3 presents the calibrated parameters. The calibrated parameters were within the ranges and details of the parameters that were described in [18] and in the SWAT user manuals. A measure of Nash Sutcliffe Efficiency (NSE) was chosen as an objective function. The NSE statistics for the monthly streamflow of the calibration and validation period were 0.84 and 0.71. More information about the calibration and validation of the SWAT model is available in [14].

3.3 Climate Change Impact on Precipitation The average yearly precipitation of Nagavali basin is 1251 mm, over the baseline period of 26 years (i.e., 1980–2005). The percentage bias between the IMD gridded precipitation and climate model precipitation for the baseline period is presented in Table 4. As shown in Table 4, the precipitation bias between the observed IMD and climate models ranged from 5.15 to 8.6%. The selected climate models represents the slightly overestimated precipitation over the basin. The percentage change between the precipitation predicted by the climate models in the near term (2022–2047), middle future (2048–2073) and far future (2074–2099) in comparison to historical data (1980–2005) is presented in Table 5. According to the table, the BNU-ESM and IPSL-CM5A-MR models revealed a decreasing precipitation in the near term and a rise in the middle and far future under the RCP 4.5 scenario. The CNRM-CM5 model predicted an increasing trend of precipitation throughout the future. For the RCP 4.5 scenario, the CNRM-CM5

Impact of Climate Change on Streamflow Over Nagavali Basin, India

305

Table 3 Calibrated parameters, range and fitted values S.No

Parameter_Name

Method

Range

Fitted values

1

CN2.mgt

Relative

±0.1

−0.088

2

ALPHA_BF.gw

Replace

0 to 1

0.642

3

GW_DELAY.gw

Absolute

-30 to 90

84.300

4

GWQMN.gw

Absolute

±1000

5

5

GW_REVAP.gw

Replace

0.02 to 0.2

0.053

6

REVAPMN.gw

Absolute

±750

−498.75

7

ALPHA_BF_D.gw

Replace

0–1

0.45

8

RCHRG_DP.gw

Absolute

−0.05 to 0.05

−0.019

9

SOL_AWC.sol

Relative

±0.1

0.04

10

ESCO.hru

Replace

0.3 to 0.6

0.53

11

CANMX.hru

Replace

0–20

0.45

12

CH_N2.rte

Replace

0.01–0.1

0.033

13

CH_K2.rte

Replace

0–100

74.75

14

CH_K1.sub

Replace

0–100

73.25

15

CH_N1.sub

Replace

0.01–0.3

0.19

Source Nagireddy et al. [14]

Table 4 Percentage bias between the IMD and historical models precipitation

Model name

% bias in precipitation

BNU-ESM

8.6

CNRM-CM5

5.15

IPSL-CM5A-MR

6.18

model showed the highest percentage change of 22.06% in the far future, and the IPSL-CM5A-MR model showed the lowest percentage change of −1.87% in the near term. Under the RCP 8.5 scenario, the IPSL-CM5A-MR model showed decreasing precipitation in the near term and increasing precipitation in the middle and far future. The BNU-ESM and CNRM-CM5 models showed an increasing trend in precipitation throughout the future. Under the RCP 8.5 scenario, the IPSL-CM5A-MR model showed the highest percentage change of 26.28% in the far future, and the lowest percentage change of −4.89% in the near future. The mean monthly precipitation and streamflow over the historical period of 1980–2005 are presented in Fig. 2. According to Fig. 2, the IMD observed precipitation peaks occurred in August with an intensity of 254.65 mm for the basin. The BNU-ESM and IPSL-CM5AMR models’ historical precipitation peaks were observed in July, but the CNRMCM5 model showed its peak precipitation in August. The projected mean monthly precipitation under two RCP scenarios is shown in Fig. 3. The BNU-ESM model projected an identical historical peak in near and mid futures, with a peak shift to August in the far future under the RCP 4.5 scenario. The

306

N. R. Nagireddy and K. Venkatareddy

Table 5 Percentage change in precipitation and streamflow with respect to historical data RCP scenario

Period

Model name

4.5

Near term (2022–2047)

BNU-ESM

−0.65

−4.92

CNRM-CM5

+8.54

+16.67

IPSL-CM5A-MR

−1.87

−9.12

Middle future (2048–2073)

Far future (2074–2099)

8.5

Near term (2022–2047)

Middle future (2048–2073)

Far future (2074–2099)

% change in precipitation

% change in streamflow

BNU-ESM

+7.03

+10.13

CNRM-CM5

+6.57

+7.82

IPSL-CM5A-MR

+17.12

+25.62

BNU-ESM

+10.78

+18.40

CNRM-CM5

+22.06

+36.26

IPSL-CM5A-MR

+13.58

+10.38 −3.27

BNU-ESM

+0.85

CNRM-CM5

+5.21

+ 7.46

IPSL-CM5A-MR

−4.89

−15.69

BNU-ESM

+21.4

+39.78

CNRM-CM5

+11.54

+11.92

IPSL-CM5A-MR

+12.54

+14.12

BNU-ESM

+22.58

+45.11

CNRM-CM5

+22.09

+33.81

IPSL-CM5A-MR

+26.28

+42.43

Note ‘+’ sign indicates increasing in future, ‘−’sign indicates decreasing in future

CNRM-CM5 model predicted peak precipitation in August at a similar historical peak. The IPSL-CM5A-MR model predicted that the peak would shift to August. The BNU-ESM model predicted a near-historical peak in near and mid futures, with a peak shift to August in the far future under the RCP 8.5 scenario. The CNRM-CM5 model predicted peak precipitation in July for the near and mid future, and August for the far future. The IPSL-CM5A-MR model predicted that the peak would shift to August.

3.4 Climate Change Impact on Streamflow According to the table, the BNU-ESM and IPSL-CM5A-MR models revealed a negative streamflow in the near term and an increase in the middle and far future under the RCP 4.5 scenario. The CNRM-CM5 model predicted an increasing trend of streamflow throughout the future. For the RCP 4.5 scenario, the CNRM-CM5 model showed the highest percentage change of 36.26% in the far future, and the IPSL-CM5A-MR

Impact of Climate Change on Streamflow Over Nagavali Basin, India

307

Fig. 2 Mean monthly precipitation and streamflow for the historical period (1980–2005)

model predicted the lowest percentage change of −9.12% in the near term. The BNUESM and IPSL-CM5A-MR model predicted decreasing streamflow in the near term and increasing streamflow in the middle and far future, under the RCP 8.5 scenario. The CNRM-CM5 model revealed an increasing trend in streamflow throughout the future. The BNU-ESM model showed the highest percentage change of 45.11% in the far future, and the IPSL-CM5A-MR model revealed the lowest percentage change of −15.69% in the near term, under the RCP 8.5 scenario. The findings are consistent with the results reported by Nagireddy et al. [10]. According to Fig. 2, the simulated streamflow peaks with IMD data occurred in August with an intensity of 39.31 mm for the basin. The BNU-ESM and IPSL-CM5A-MR models’ historical streamflow peaks were observed in July, but the CNRM-CM5 model showed its peak streamflow in August. The projected mean monthly streamflow under two RCP scenarios is shown in Fig. 4. The BNU-ESM model projected a similar historical peak in the near and middle future, but a peak shift to August in the far future, under the RCP 4.5 scenario. The CNRM-CM5 model predicted the peak streamflow in August at a similar historical peak. The IPSL-CM5A-MR model predicted that the peak would shift to August. Under the RCP 8.5 scenario, the BNU-ESM model predicted a similar historical

308

N. R. Nagireddy and K. Venkatareddy

Fig. 3 Projected mean monthly precipitation under two RCP scenarios

peak in the near and mid future, but a peak shift to August in the far future. The CNRM-CM5 model predicted the peak precipitation in July for the near and mid future, and August for the far future. The IPSL-CM5A-MR model predicted that the peak would shift to August.

4 Conclusions Flood frequency assessment using log Pearson type III distribution shows more frequent low magnitude floods and longer period of record will be critical for a complete evaluation for designing structures such dam, bridges, culverts etc. The percentage bias between the IMD precipitation and climate models precipitation over the basin represents slightly overestimation. The percentage change between

Impact of Climate Change on Streamflow Over Nagavali Basin, India

309

Fig. 4 Projected mean monthly streamflow under two RCP scenarios

the historical and future simulated streamflow patterns followed the climate model precipitation patterns. The BNU-ESM and IPSL-CM5A-MR models revealed a negative precipitation and streamflow over near term and a positive over the middle and far future. The CNRM-CM5 model predicted an increasing trend of precipitation and streamflow throughout the future under both RCP scenarios. The IMD observed precipitation and simulated streamflow peaks occurred in August. The BNU-ESM and IPSL-CM5A-MR models’ historical precipitation and streamflow peaks were observed in July, but the CNRM-CM5 model showed its peak precipitation and streamflow in August. The BNU-ESM model precipitation and streamflow showed a similar historical peak in the near and mid future, but a peak shift to August in the far future, under the RCP 4.5 scenario. The CNRM-CM5 model predicted the peak precipitation and streamflow in August similar to historical peak. The IPSLCM5A-MR model precipitation and streamflow predicted that the peak would shift to August. The BNU-ESM model projected a similar historical peak in the near and

310

N. R. Nagireddy and K. Venkatareddy

mid future, but a peak shift to August in the far future, under the RCP 8.5 scenario. The CNRM-CM5 model predicted the peak precipitation and streamflow in July for the near and mid future, and August for the far future. The IPSL-CM5A-MR model predicted that the peak would shift to August.

References 1. Abbaspour KC, Faramarzi M, Ghasemi SS, Yang H (2009) Assessing the impact of climate change on water resources in Iran. Water Resour Res 45(10) 2. Wagner PD, Bhallamudi SM, Narasimhan B, Kantakumar LN, Sudheer KP, Kumar S, Schneider K, Fiener P (2016) Dynamic integration of land use changes in a hydrologic assessment of a rapidly developing Indian catchment. Sci Total Environ 539:153–164 3. Dhar S, Mazumdar A (2009) Hydrological modelling of the Kangsabati river under changed climate scenario: case study in India. Hydrol Process: Int J 23(16):2394–2406 4. Gosain AK, Rao S, Basuray D (2006) Climate change impact assessment on hydrology of Indian river basins. Curr Sci 346–353 5. Gosain, A. K., Sandhya Rao., Anamika Arora.: Climate change impact assessment of water resources of India. Current Science (2011): 356–371. 6. Menon A, Levermann A, Schewe J (2013) Enhanced future variability during India’s rainy season. Geophys Res Lett 40(12):3242–3247 7. Choudhury BA, Rajesh PV, Zahan Y, Goswami BN (2021) Evolution of the Indian summer monsoon rainfall simulations from CMIP3 to CMIP6 models. Clim Dyn 1–26 8. Raghavan SV, Hur J, Liong SY (2018) Evaluations of NASA NEX-GDDP data over Southeast Asia: present and future climates. Clim Change 148: 503–518 9. Rao KK, Kulkarni A, Patwardhan S, Kumar BV, Kumar TL (2020) Future changes in precipitation extremes during northeast monsoon over south peninsular India. Theoret Appl Climatol 142:205–217 10. Nagireddy NR, Keesara VR, Venkata Rao G, Sridhar V, Srinivasan R (2023) Assessment of the impact of climate change on streamflow and sediment in the Nagavali and Vamsadhara Watersheds in India. Appl Sci 13(13):7554 11. Griffis VW, Stedinger JR (2007) Log-Pearson type 3 distribution and its application in flood frequency analysis. I: distribution characteristics. J Hydrol Eng 12(5):482–491 12. Pawar U, Hire P (2018) Flood frequency analysis of the Mahi Basin by using Log Pearson Type III probability distribution. Hydrospatial Anal 2(2):102–112 13. Bhat MS, Alam A, Ahmad B, Kotlia BS, Farooq H, Taloor AK, Ahmad S (2019) Flood frequency analysis of river Jhelum in Kashmir basin. Quatern Int 507:288–294 14. Nagireddy NR, Keesara VR, Sridhar V, Srinivasan R (2022) Streamflow and sediment yield analysis of two medium-sized east-flowing river basins of India. Water 14(19):2960 15. Neitsch SL, Arnold JG, Kiniry JR, Williams JR (2011) Soil and water assessment tool theoretical documentation version 2009. Texas Water Resources Institute 16. Nash JE, Sutcliffe JV (1970) River flow forecasting through conceptual models part I—a discussion of principles. J Hydrol 10(3):282–290 17. Moriasi DN, Arnold JG, Van Liew MW, Bingner RL, Harmel RD, Veith TL (2007) Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans ASABE 50(3):885–900 18. Arnold JG, Moriasi DN, Gassman PW, Abbaspour KC, White MJ, Srinivasan R, Santhi C et al (2012) SWAT: model use, calibration, and validation. Trans ASABE 55(4):1491–1508

Impervious Surface Area Prediction Using Landsat Satellite Imagery and Open Source GIS Plugin Ayyappa Reddy Allu and Shashi Mesapam

Abstract In India, population growth is rapidly increasing and the rural people are moving to urban areas for improving their socio-economic activity of life. For sustaining the human needs, most of the land use land cover features are migrating to impervious surfaces, which may lead to decreasing the infiltration capacity of the soil and increasing the flood frequency. Land Use Land Cover (LULC) maps are helpful to monitor and predict the impervious surface area using the Remote Sensing techniques. The proposed work aims to simulate the impervious surface area of Jagtial, Telangana, India, in the year 2050. This is accomplished by using Landsat satellite images from the years 2000, 2005, 2010, 2015, and 2020, and applying the Random Forest classification algorithm to generate LULC maps. The maximum tree depth is set at 30, the maximum number of trees is 250, and the maximum number of samples per class is 1000. A land use simulation model, based on Cellular Automata and Markov Chains, is employed to calibrate and optimize the LULC images. The model predicts the LULC map for the years 2020 and 2050, which are validated using existing classified LULC images. The imperviousness index of the LULC classes is used to estimate the impervious surface area of the location. The analysis of the multi-temporal LULC images shows that biophysical and socioeconomic factors have a significant impact on the increase in built-up areas and the decline in water bodies by the year 2050. Keywords Cellular automata · Impervious surface area · Land use land cover · Markov chain

1 Introduction In urban areas, population growth is rapidly growing due to the increase in population and migration of the people from rural places. LULC classes are transforming to the built up and industrial areas for meeting the requirements of the increased population, A. R. Allu (B) · S. Mesapam Department of Civil Engineering, National Institute of Technology, Warangal, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_23

311

312

A. R. Allu and S. Mesapam

which may lead to increase in the impervious surface. Impervious surface is one of the land cover form which may prevents the infiltration of water into the soil and it may also impact the ground water table and increasing the runoff [1, 2]. Human made surfaces such as pavements, buildings, parking lots, are constructed with the concrete, whereas the asphalt material are considered as an impervious surface [1, 3]. Impervious surface area will be playing a major role in urban planning and development, and resource management [1, 3, 4]. So, there is a need to monitor and predict the impervious surfaces to regularize the development. Remote Sensing is one of the technology which uses the satellite images to continuous assessment and monitoring of impervious surfaces [5]. Several techniques were successfully applied to extract the impervious surface location using the satellite images in the previous research works, including Artificial Neural Network (ANN), sub pixel classification, regression tree analysis, multiple regression analysis, spectral mixture analysis, LULC classification [3, 5–9] etc. In that, the LULC classification method is one of the easiest method to solving the environmental problem [10, 11]. LULC images are playing a vital role in the identification of impervious surfaces by classifying the land use features on the ground into number of classes [12]. Each class has different impervious surface characteristics, that will be helpful for extracting the impervious surface area [10, 13]. Classifying the features into classes in the satellite images might be a challenging task due to their spatial and spectral information. Training dataset will be helpful for classifying the satellite images by training the random sample of each class from the respective satellite image [12]. Many classification algorithms, such as Maximum Likelihood, Support Vector Machine, and Random Forest, have been shown to be useful for classifying the features to prepare the Land Use and Land Cover (LULC) maps using training datasets [4, 9, 11, 14–16]. Each classification algorithm has its own strengths and limitations. The Random Forest classification algorithm, in particular, uses an ensemble of multiple decision trees to make decisions, resulting in high accuracy in classifying the features into the relevant classes [3, 6, 8, 15, 17–20]. This algorithm builds a decision tree for each sample to train the dataset. The combination of impervious nature of LULC classes provides the impervious surface location [1, 2, 8, 21] of the respective satellite image but for prediction, it needs to simulate the future image of the LULC by using the existing dataset. Modelling is an effective method to provide the opportunity to explore multiple pathways to analyses and predict the future from temporal data [10, 19]. Several studies worked in the modelling of LULC change to simulate land use changes and develop the interaction between land cover and environmental changes [9, 14, 15, 19]. LULC changes mainly affect the biodiversity, vegetation, water resources, and human life, and are linked to sustainable economic development [4]. LULC change analysis in most of the studies is worked with the regression model or transition model to simulate and predict the future datasets. Spatial transition models simulate the future development with possible estimations using the Monte Carlo’s based Cellular Automata (CA) and Markov Chain (MC) or other methods [11, 16, 17]. The CA-MC is an integrated model which is suitable to detect the land use changes and simulates to predict the land cover changes by considering the spatial

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

313

and temporal components of the dynamic change in land cover [14]. This CA-MC model is included in various software such as TerraSet, QGIS [14]. In that, QGIS is an open source software which uses the MOLUSCE plugin [4, 12, 14, 16, 17, 22] (Modules for Land Use Change Evaluation) (https://plugins.qgis.org/plugins/ molusce/) to simulate the future images from the past LULC images using the ANN, Logistic Regression (LR), Weight of Evidence and Multi Criteria Evaluation techniques [16, 17, 19]. For these techniques, transitional area change of LULC classes plays a major role for modelling and simulating the future developments and changes [22]. The periodic gap between the LULC images is also an impacting factor for predicting the LULC changes but it depends on the growth of the study area location. The primary objective of this study is to predict the extent of impervious surfaces in the year 2050 using Landsat satellite images from the years 2000, 2005, 2010, 2015, and 2020 on the QGIS platform. LULC classification has been prepared from the Landsat satellite images to extract the impervious surface area of the study area location in QGIS software. The future LULC image will be predicted using the MOLUSCE plugin in the QGIS software. The novelty of this approach is to predict the impervious surface of the study area by utilizing different periodic change combination of the initial and final year satellite images.

2 Study Area Jagtial district is located in the state of Telangana, India (refer to Fig. 1). It is situated between the latitude 18°42’ and longitude 78°54’ and has an average elevation of +293 m above the sea level. According to the 2011 census, the district had a population of 9,85,417 and a total area of 2,419 KM2 . Jagtial district is divided into 5 municipalities, 18 mandals, and 380 villages. Industries, educational institutions, and research institutes are located in the Jagtial and Korutla municipalities, which have led to significant changes in these regions.

3 Methodology The methodology flowchart (refer to Fig. 2) outlines the steps involved in data collection, preparation, processing, analysis, and validation to estimate and simulate the impervious surface area for the year 2050 using the Landsat images.

3.1 Data Collection In this study, Landsat satellite imagery from the years 2000, 2005, 2010, 2015, and 2020 were utilized to estimate and predict the impervious surface area of the study

314

A. R. Allu and S. Mesapam

Fig. 1 Jagtial location in India map

Fig. 2 Methodology flowchart of impervious surface map using Landsat satellite images

location. The satellite imagery from the years 2000, 2005, and 2010 was obtained from the Landsat 05 TM (Thematic Mapper), while the imagery from 2015 and 2020 was obtained from the Landsat 8 OLI (Operational Land Imager) through GloVis USGS (https://glovis.usgs.gov/). The blue, green, red, and Near Infra-Red (NIR) spectral bands of the Landsat satellite imagery were utilized with a spatial resolution of 30 m. The satellite images acquired from Landsat were all in the same reference system (WGS 1984 UTM Zone 44N), and the information regarding the collected data is presented in Table 1. For validation purposes, Google Earth, field validation, and socio-economic activities were considered as ground truth data.

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

315

Table 1 Data acquisition information about the Landsat satellite imagery Data

Path

Row

Date of collection

Bands

Landsat 05 TM

144

47

02 October 2000

1, 2, 3, 4

Landsat 05 TM

144

47

07 April 2005

1, 2, 3, 4

Landsat 05 TM

144

47

23 May 2010

1, 2, 3, 4

Landsat 08 OLI

144

47

19 April 2015

2, 3, 4, 5

Landsat 08 OLI

144

47

11 January 2020

2, 3, 4, 5

3.2 Data Preparation Preprocessing techniques were applied to the Landsat images to perform atmospheric correction. The preprocessed bands of the satellite imagery for the years 2000, 2005, 2010, 2015, and 2020 were layer stacked, and the study area location was extracted from the layer-stacked image for the respective years by using its administrative boundary (shown in Fig. 3). Subsequently, image classification techniques were performed for further processing and analysis. Image Classification The Random Forest (RF) algorithm was employed to classify the layer-stacked images. The samples of each class for every year were used to train the dataset. The area was classified into four (LULC) classes: barren, built-up, vegetation, and water bodies, which represented the features in the satellite image (Table 2). For

Fig. 3 Layer stacked images of Jagtial location for the year a 2000, b 2005, c 2010, d 2015 and e 2020

316

A. R. Allu and S. Mesapam

each class, 30 training samples were selected, and the satellite images were classified using the Random Forest algorithm with a maximum tree depth of 30, a maximum number of trees of 250, and a maximum number of samples per class of 1000. The raster images were generated in the open-source GIS software, QGIS (https://www.qgis.org/en/site/), and the satellite images were classified using the Semi-Automatic Classification (https://github.com/semiautomaticgit/SemiAutomati cClassificationPlugin) plugin. The LULC classification was performed for the years of 2000, 2005, 2010, 2015 and 2020 using the RF supervised machine learning algorithm. The results of the classification are presented in Fig. 4.

Table 2 Land cover feature of the LULC class S. No

Class

Feature

1

Barren

Bare lands, sediment deposits, exposed rocks, mixed grasslands

2

Builtup

Residential, commercial, industrial, roads, railway network, and other uncategorized surface areas

3

Vegetation

Agriculture, forest grass land, shrubs and bush areas

4

Water Body

Open water bodies like ponds, lakes, rivers, reservoirs

Fig. 4 LULC images of the year: a 2000, b2005, c 2010, d 2015 and e 2050

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

317

Table 3 Accuracy of the classified images for the years of 2000, 2005, 2010, 2015 and 2020 Year of LULC image

2000

2005

2010

2015

2020

Overall accuracy (%)

86.05

78.17

82.15

81.73

80.54

Kappa coefficient (Kc )

0.795

0.744

0.781

0.797

0.760

Accuracy Assessment The accuracy of the classified LULC images was validated using the ground truth information collected from Google Earth and field. A confusion matrix was created to evaluate the accuracy of the classified images, which was measured through the overall accuracy and Kappa coefficient. The equations used for this evaluation are as follows: Number of pixels classified correctly × 100 Total number of pixels   N ri=1 xi j − ri=1 (x i+ × x+ j )  Kappa Coefficient(K c ) = N 2 − ri=1 (x i+ × x+ j )

Overall Accuracy(O A) =

where, x ij is the number of observations in row i and column j, x i+ is the marginal total of row i, x +j is the marginal total of column j, r is the total number of rows and columns in the error matrix and N is the total number of observations. The range of the Kappa coefficient (K c ) typically falls between −1 and +1, but it is usually between 0 and 1. A value of 1 indicates complete agreement between the classified and reference data but zero represents the no agreement. The accuracy of the classified LULC images (Table 3) is useful for calculating the impervious surface area and predicting future data.

3.3 Data Processing Impervious Surface Ratio The Impervious Surface Ratio (ISR) of the study location indicates the amount of land covered with hard materials that resist water infiltration. The LULC images were used to evaluate and predict the impervious surface by monitoring changes in the natural landscape. In this study, the features were classified into four classes, with the vegetation, barren, and water body classes considered as pervious surfaces and the built-up class considered as the impervious surface. The ISR was calculated using the following equation: Impervious Surface Ratio =

Total Impervious Surface Area Total(Pervious + Impervious)Area

318

A. R. Allu and S. Mesapam

Fig. 5 Impervious surface map for the year a 2000, b 2005, c 2010, d 2015 and e 2020

ISR maps (Fig. 5) for the years 2000, 2005, 2010, 2015 and 2020 were prepared using the classified LULC images from those respective years. Simulated LULC images will be useful for predicting the ISR for the year of 2050. Simulation of LULC using MOLUSCE This study predicts the LULC for the year of 2050 through simulation of three different combinations of the initial and final year images: (i) the LULC of 2000 and 2005, (ii) the LULC of 2000 and 2010, and (iii) LULC of 2010 and 2015. CA based MC calibration and simulation model was used to predict the future LULC image from past LULC maps. The model predicts the transition from one state at T + t to T + n(t) state based on the state T. L (T +t) = Pi j × L T ⎡

⎤ P11 P12 P13 Pi j = ⎣ P21 P22 P23 ⎦ P31 P32 P33 where L (T ) is the land use status at the year of T, L (T +t) is the land use at the year of T + t, T is the initial year, t is the change in time, n is number of times of simulation and Pi j is the i, jth cell of contingency table.

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

319

The MOLUSCE plugin was utilized to predict the LULC of the year 2050 through the simulation of LULC changes between different periodic combinations. The functional procedure of the MOLUSCE plugin is described as follows: Correlation Evaluation A correlation analysis was conducted to evaluate the linear relationship between the initial and final year LULC images. This was accomplished by assessing the statistical measures of the spatial variable factors using the Pearson’s correlation coefficient. Area Change Transition area and transition probability matrices were created between the initial and final year LULC images to analyze the trends in the transformation of LULC features. The change in area (KM2 ) and probability of area change (%) from one class to another classes were measured in the transition matrix. Transitional Potential Modelling Modeling the variation of LULC features and recognizing the patterns is a challenging task for human brains, but algorithms can be used to model complex patterns. Fuzzy logic requirements were incorporated into an ANN, to allow for a continuous range between 0 and 1, determined by the usability of the terrain. The core elements of the ANN inter-connected between the neurons, with modifications to the weight connections were used to learn the patterns of transforming direction of LULC features. Cellular Automata Simulation CA is used to simulate the LULC by taking the transition probabilities from the transition potential modelling. Transition simulator maps, certaincy and simulation maps were generated for the each simulation. Validation The predicted LULC images were validated using existing datasets to predict future LULC images. The validation module used four parameters of kappa statistics to calculate the accuracy of the model: kappa location (K loc ), kappa histogram (K hist ), overall kappa (K) and percentage of correctness. K loc =

P(A) − P(E) Pmax − P(E)

K hist =

Pmax − P(E) 1 − P(E)

K = K loc × K hist =

P( A) − P(E) 1 − P(E)

320

A. R. Allu and S. Mesapam

where c c c P( A) = i=1 pii , P(E) = i=1 ( pia × pa j) , Pmax = i=1 min( pia , pa j) , pij is the i, jth cell of contingency table, pia is the sum of all cells in ith row, paj is the sum of all cells in jth column, and c is count of raster categories.

4 Results and Discussion The changes in LULC features are not constant over time and may be influenced by various factors, such as socio-economic activities, traffic, etc. The analysis of the classified LULC images shows that barren land areas have changed to vegetation and built-up areas from 2005 to 2010 and 2015, while vegetation has changed to built-up and barren land. However, the built-up area has increased from 2000 to 2020. Table 5 shows the changes in the area of classified features for the years 2000 to 2020, which can be used to predict future LULC images. In this study, simulation techniques are used to predict the LULC for the year 2050 by validating the predicted LULC of 2020 with the classified LULC of 2020. The Pearson’s correlation coefficient between the input images is high (as shown in Table 4), indicating a strong relationship between the LULC images. A transition probability matrix was created between the initial and final year LULC images, representing the probability of each feature class changing from one class to another. Table 5 shows that the transformation of water body features to built-up is high compared to the other classes. Vegetation features are mostly transforming to barren and built-up classes in every combination, while barren classes are transforming to built-up at a much lower rate from 2010 to 2015 compared to other combinations. On the other hand, most of the built-up features are transforming to vegetation in the period between 2000 and 2010, and between 2010 and 2015, compared to the period between 2000 and 2005. This suggests that there was an increase in vegetation and a decrease in the rate of change of built-up area from 2010 to 2015, with most of the built-up area being constructed in 2005. The MC-CA technique used an Artificial Neural Network (ANN) algorithm to predict the future LULC images for 2020 and validated the results using the classified LULC image of that year. If the accuracy of the model was validated, it could then be used to predict the LULC image for 2050. Figure 6 shows the simulated LULC images for the years 2020 and 2050, created using a combination of initial and final images from 2000 and 2005, 2000 and 2010, and 2010 and 2015. The accuracy of the simulated images is presented in Table 6. Table 4 Pearson’s correlation coefficient between initial and final year images

Initial image

2000

2000

2010

Final image

2005

2010

2015

Perason’s coefficient

0.91524

0.93101

0.94427

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

321

Table 5 Transition probability matrix between initial and final year 2000–2005

Water body

Builtup

Vegetation

Barren

Water body

0.182212

0.433089

0.188618

0.196082

Builtup

0.001585

0.51499

0.296325

0.1871

Vegetation

0.000463

0.173612

0.519008

0.306916

Barren

0.000161

0.178949

0.345664

0.475226

Water body

0.194328

0.305635

0.375121

0.124916

Builtup

0.001131

0.444525

0.426566

0.127779

Vegetation

0.000911

0.112287

0.679013

0.207789

Barren

0.000186

0.110661

0.474718

0.414435

Water body

0.650648

0.230543

0.056886

0.061923

Builtup

0.021064

0.395258

0.485827

0.097851

Vegetation

0.002802

0.193037

0.655715

0.148446

Barren

0.005275

0.056041

0.249508

0.689176

2000–2010

2010–2015

Fig. 6 Simulated LULC images: a, c, e represent the simulated year of 2020, created using the initial and final images of 2000, 2005; 2000, 2010; and 2010, 2015 respectively. b, d, f represent the simulated year of 2050, created using the same combinations of initial and final images

322

A. R. Allu and S. Mesapam

Table 6 Accuracy of simulated LULC images of the year 2020 and 2050 using the initial and final year images Initial

Final

Simulated

Kappa Histo

Kappa local

Kappa

% of correctness

2000

2005

2020

0.62369

0.80497

0.77479

73.19752

2000

2005

2050

0.63475

0.75241

0.72386

71.25894

2000

2010

2020

0.84649

0.91443

0.77406

85.38166

2000

2010

2050

0.76889

0.77242

0.69391

78.64269

2010

2015

2020

0.80482

0.71247

0.57341

69.25772

2010

2015

2050

0.85098

0.69286

0.58961

70.68681

The simulated LULC images for the year 2020 and 2050 are accurately predicted using the initial and final images of 2000 and 2010. These images depict an increase in vegetation and built-up areas and a decrease in barren and water body areas when compared to the latest year LULC image (2020). The most significant increase in built-up area was observed in the west side (Ibrahimpatnam, Metpalli, Kathlapur, Mallapur, and Korutla mandals) and the north-east side (Dharmapuri and Velgatur mandals) of the study area. However, the percentage area of these feature classes in both the classified and simulated LULC images is shown in Table 7. These results demonstrate the potential of the simulation techniques and the power of the ANN algorithm in predicting future LULC images and understanding the impact of socioeconomic activities and environmental changes on the study area. Figure 5 indicates that, there has been a limited increase in the impervious surface ratio in the study area. The results of the study show that the increase in built-up areas is concentrated in the west and north-east regions of Jagtial district, while the central region of the district has remained largely unaffected (shown in Fig. 7). This suggests that the impact of socio-economic activities and environmental changes on the study area has been relatively limited, at least with regards to the impervious surface ratio. The results of the study may be useful for decision-makers, planners and other stakeholders in the area to develop strategies for managing and mitigating the impact of future growth and development in the region.

Table 7 Percentage area of the various classes in classified and simulated LULC images Class/year Classified area (%)

2000

2005

2010

2015

2020

Simulated area from 2000 and 2005 (%)

Simulated area from 2000 and 2010 (%)

Simulated area from 2010 and 2015 (%)

2020

2020

2020

2050

2050

2050

Barren

34.90 35.78 27.44 28.97 40.71 35.94 56.63 22.84 21.91 31.31 35.24

Builtup

4.24

19.65 13.07 18.21 13.57 19.95 7.70

16.48 18.83 23.37 24.28

Vegetation 58.33 44.07 58.93 51.87 43.42 43.74 35.23 60.27 59.07 44.39 39.53 Water Body

2.53

0.50

0.56

0.95

2.29

0.37

0.44

0.41

0.19

0.94

0.95

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

323

Fig. 7 Impervious Surface Ratio images of the simulated LULC images: a, c, e are simulated year of 2020 from the initial and final images of 2000, 2005; 2000, 2010; and 2010, 2015 respectively, b, d, f are simulated year of 2050 from the initial and final images of 2000, 2005; 2000

5 Conclusion The impervious surface map for 2050 was predicted using open-source GIS software. The prediction was based on the classified LULC images from the years 2000, 2005, 2010, 2015, and 2020, which were processed using the MOLUSCE plugin in QGIS software. The results of the prediction show an increase in built-up areas in the west and northeast regions, a conversion of water bodies into built-up areas, and negligible changes in vegetation in the central part of the district. The impervious surface ratio in the study area does not change significantly. The LULC changes in the study area appear to vary linearly. QGIS played a crucial role in this study, helping to address the problems arising from socio-economic activities and environmental changes.

References 1. Majid MR, Jamaludin JA, Ibrahim WYW (2013) Estimation of residential impervious surface using GIS technique. Plan Malaysia 11:23–38. https://doi.org/10.21837/pmjournal.v11.i2.114 2. Ongsomwang S, Pimjai M (2014) Land use and land cover prediction and its impact on surface runoff 3. Zhang H, Gorelick SM, Zimba PV (2020) Extracting impervious surface from aerial imagery using semi-automatic sampling and spectral stability. Remote Sens 12(3). https://doi.org/10. 3390/rs12030506

324

A. R. Allu and S. Mesapam

4. Kamaraj M, Rangarajan S (2022) Predicting the future land use and land cover changes for Bhavani basin, Tamil Nadu, India, using QGIS MOLUSCE plugin. Environ Sci Pollut Res 29(57):86337–86348. https://doi.org/10.1007/s11356-021-17904-6 5. Hu X, Weng Q (2009) Estimating impervious surfaces from medium spatial resolution imagery using the self-organizing map and multi-layer perceptron neural networks. Remote Sens Environ 113(10):2089–2102. https://doi.org/10.1016/j.rse.2009.05.014 6. Feng S, Fan F (2021) Impervious surface extraction based on different methods from multiple spatial resolution images: a comprehensive comparison. Int. J. Digit. Earth 14(9):1148–1174. https://doi.org/10.1080/17538947.2021.1936227 7. Kaspersen PS, Fensholt R, Drews M (2015) Using Landsat vegetation indices to estimate impervious surface fractions for European cities. Remote Sens 7(6):8224–8249. https://doi. org/10.3390/rs70608224 8. Zhang X et al (2020) Development of a global 30m impervious surface map using multisource and multitemporal remote sensing datasets with the Google Earth Engine platform. Earth Syst. Sci. Data 12(3):1625–1648. https://doi.org/10.5194/essd-12-1625-2020 9. Lu D, Moran E, Hetrick S (2011) Detection of impervious surface change with multitemporal Landsat images in an urban-rural frontier. ISPRS J Photogramm Remote Sens 66(3):298–306. https://doi.org/10.1016/j.isprsjprs.2010.10.010 10. Li W, Wu C, Choi W (2018) Predicting future urban impervious surface distribution using cellular automata and regression analysis. Earth Sci Inform 11(1):19–29. https://doi.org/10. 1007/s12145-017-0312-8 11. Al Kafy A et al (2021) Predicting changes in land use/land cover and seasonal land surface temperature using multi-temporal landsat images in the northwest region of Bangladesh. Heliyon 7(7). https://doi.org/10.1016/j.heliyon.2021.e07623 12. Satya BA, Shashi M, Deva P (2020) Future land use land cover scenario simulation using open source GIS for the city of Warangal, Telangana, India. Appl Geomatics 12:281–290. https:// doi.org/10.1007/s12518-020-00298-4/Published 13. Nair HM (2013) Estimation of effective impervious surface area of cochin using satellite images. [Online]. www.isca.in 14. Rahnama MR (2021) Forecasting land-use changes in Mashhad metropolitan area using cellular automata and Markov chain model for 2016–2030. Sustain Cities Soc 64. Elsevier Ltd. https:// doi.org/10.1016/j.scs.2020.102548 15. Liping C, Yujun S, Saeed S (2018) Monitoring and predicting land use and land cover changes using remote sensing and GIS techniques—a case study of a hilly area, Jiangle, China. PLoS One 13(7). https://doi.org/10.1371/journal.pone.0200493 16. Muhammad R, Zhang W, Abbas Z, Guo F, Gwiazdzinski L (2022) Spatiotemporal change analysis and prediction of future land use and land cover changes using QGIS MOLUSCE plugin and remote sensing big data: a case study of Linyi, China. Land 11(3). https://doi.org/ 10.3390/land11030419 17. Alshari EA, Gawali BW (2022) Modeling land use change in Sana’a city of Yemen with MOLUSCE. J Sensors 2022. https://doi.org/10.1155/2022/7419031 18. Parekh JR, Poortinga A, Bhandari B, Mayer T, Saah D, Chishtie F (2021) Automatic detection of impervious surfaces from remotely sensed data using deep learning. Remote Sens 13(16). https://doi.org/10.3390/rs13163166 19. Kafy AA et al (2021) Cellular automata approach in dynamic modelling of land cover changes using RapidEye images in Dhaka, Bangladesh. Environ. Challenges 4. https://doi.org/10.1016/ j.envc.2021.100084 20. Li F et al (2021) Estimating artificial impervious surface percentage in asia by fusing multitemporal modis and viirs nighttime light data. Remote Sens 13(2):1–23. https://doi.org/10. 3390/rs13020212 21. Das N, Mondal P, Sutradhar S, Ghosh R (2021) Assessment of variation of land use/land cover and its impact on land surface temperature of Asansol subdivision. Egypt J Remote Sens Sp Sci 24(1):131–149. https://doi.org/10.1016/j.ejrs.2020.05.001

Impervious Surface Area Prediction Using Landsat Satellite Imagery …

325

22. Sajan B, Mishra VN, Kanga, Meraj G, Singh SK, Kumar P (2022) Cellular automata-based artificial neural network model for assessing past, present, and future land use/land cover dynamics. Agronomy 12(11). https://doi.org/10.3390/agronomy12112772

Influence on Water Characteristics Away from Various Sources of NIT Kurukshetra Using ArcGIS Rahul Deopa and K. K. Singh

Abstract In this paper we assess the characteristics of various water sources of the NIT Kurukshetra campus using ArcMap 10.8 Software. Water Quality Index (WQI) is used to evaluate the drinking water quality from different sources on the NIT Kurukshetra Campus using the Weighted Arithmetic Method. A way of rating water quality is the Water Quality Index. It was done to determine the area’s overall groundwater quality significance. It is a useful instrument for determining the quality of groundwater. Electroconductivity, color, odor, taste, pH, Chlorides (Cl), total dissolved solids (TDS), total alkalinity (TA), and total hardness (TH) are used in this study to calculate the quality of water at NIT Kurukshetra Campus. Water quality measurements were gathered at five distinct places on the NIT Kurukshetra Campus. The results revealed that all of the stations have Excellent water quality, according to the values of the WQI although the readings of most of the sources measured are above acceptable and occasionally permissible limits also but the water quality still comes out as excellent. The data from the evaluation was then processed using ArcGIS’ spatial analysis function in which the Kriging technique of Interpolation is used from which various spatial distribution maps for various water source locations were constructed after selecting the appropriate interpolation strategy. Using Image Overlapping, the variation of physicochemical characteristics of some parameters with length is also determined. Keywords Water quality · WQI · Kriging · Interpolation · ArcMap 10.8

1 Introduction Water is a transparent, odorless, colorless liquid that forms the lakes, seas, ponds, rivers, and rain and is the basic requirement of living organisms and also a vital source of water supply [1]. Water covers 71% of the earth’s surface. On Earth, 96.5% of the planet’s crust water is found in seas and oceans which are saline, 1.7% in groundwater, R. Deopa (B) · K. K. Singh Department of Civil Engineering, NIT Kurukshetra, Haryana, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_24

327

328

R. Deopa and K. K. Singh

1.7% in glaciers and ice caps of Antarctica and Greenland, and less than 0.3% of all freshwaters is in rivers, lakes, and ponds. All this water can be optimally sustained only when the quantity and quality are properly assessed. For healthy living, drinkable safe water is very necessary. An adequate quantity of clean and fresh drinking water is a basic human necessity. The water quality index method is one of the most effective methods to determine the water quality in which water quality is accessed with the help of water quality indices [2]. The water quality index method is one of the most effective methods for determining water quality. Water quality indices are used to access water quality. Water quality is determined by a variety of physical, chemical, and biological factors [3]. Horton in 1965 was the first to attempt to combine the several characteristics that characterize the status of water into a single value known as the Water Quality Index [4]. To create his index, Horton chose ten of the most often measured variables. He chose rating scales for each variable that range from 0 to 100, with 100 representing the highest quality rate. The weighing parameters vary between 1 and 4. The weighted sum of sub-indices is divided by the sum of weights, and two coefficients based on water temperature and pollution level are multiplied by the final index score.

2 Study Area Kurukshetra, considered one of the most religious and holy cities of India with immense reference to mythological and ancient India, is located in the state of Haryana. NIT Kurukshetra is one of the Prestigious Premier Technical Institutes situated in Kurukshetra, Haryana. Geographically, it has its locations defined at a latitude of 29o 56' 56.616'' and longitude of 79o 49' 2.114'' . The campus extends over the area of 300 acres laid down on an attractive landscape. It is impressive in architecture and natural beauty (Fig. 1).

3 Materials and Methods In the presented research work, the NIT Kurukshetra Campus is selected to assess the drinking water quality at various points of the source. There are 5 points of water source which are taken for the sampling method which are located in the different regions of the campus so that they can cover the whole campus. These sampling stations are: Tubewell 6, Tubewell 8, Tubewell 7, Tubewell 9 and Academic Section. The samples are collected from each location in December 2021 and tests are performed and analyzed in the laboratory. Color, odor, taste, electroconductivity, total hardness, total alkalinity, pH, salinity, chlorides, and total dissolved solids are among the criteria examined (TDS). The geographical distribution map of various

Influence on Water Characteristics Away from Various Sources of NIT …

329

Fig. 1 Study area (NIT Kurukshetra)

physicochemical parameters is likewise created in ArcGIS utilizing GPS coordinates and interpolation techniques in the software’s Spatial Analysis tool for various specified places (Table 1 and Fig. 2). Water Quality Index (WQI) Method This method is used to connect a group of water quality parameters and integrate them into a single number in accordance with a selected method of computation [5]. It represents the complex data in a simple and understandable way and provides a common framework for comparing a wide range of measured data with prescribed standard limits. The steps to find out WQI are as follows: Table 1 Sampling stations of NIT Kurukshetra

Location L1

Tubewell No.6

Location L2

Tubewell No.9

Location L3

Tubewell No.8

Location L4

Tubewell No.7

Location L5

Academic section

330

R. Deopa and K. K. Singh

Fig. 2 Sampling stations of NIT Kurukshetra

• • • •

Selection of parameter Determination of sub-index of the quality parameter Determination of Weightage factor Integration of sub-Indices in a mathematical expression.

Sub Index of Quality Parameter (qn ): qn = ((Vn − Vi ) / (Sn − Vi ))x100 where, V n = Observed Actual Value S n = Standard Value V i = Ideal Value Note: Except for DO, pH, and temperature, all Ideal Values for water are set to zero. Weightage Factor (W): Wn = (K /Sn ) where,

Influence on Water Characteristics Away from Various Sources of NIT …

331

Sn = (1/Σ(1/S n )). K = Proportionality Constant. Note: ΣW = 1. Water Quality Index (WQI): WQI = ((Σwi qi )/(Σwi ))

4 Results and Discussion The values of various physicochemical parameters at various sampling stations (L1 , L2 , L3 , L4 , L5 ) are given in Table 2. Various drinking water standards are given by the IS 10500:2012 [6]. Color: Color in water is generally due to dissolved organic matter from decaying vegetation or some inorganic materials lsuch as colored soil etc. 1 TCU is equal to the color produced by a 1 mg/L of platinum in the form of chloroplatinate ion. Taste and Odor: Odor is measured by an instrument called Osmoscope. Taste and Odor are caused by dissolved gasses. Algae secrete oily substances which may result in bad taste and odor. Electroconductivity: Electrical conductivity (EC) is related to total dissolved solids (TDS). The significance of EC and TDS resides in their impact on a water sample’s corrosivity as well as the solubility of barely soluble molecules like CaCO3.The conductivity increases as the ion concentration grows. It’s one of the most essential considerations when deciding whether or not water is suitable for irrigation. According to WHO recommendations, the EC value should not exceed 400 S/cm. According to the current investigation, the minimum EC value was 1373 S/cm, with a maximum of 3548 S/cm. Total Dissolved Solids: TDS is made up of dissolved inorganic salts (mostly sodium, magnesium, bicarbonates, potassium, calcium, chlorides, and sulfates) as well as minor amounts of organic stuff. TDS concentrations in water vary significantly across geological locations due to changes in mineral solubilities (WHO (World Health Organization)) [7]. A high TDS value in water indicates that it has been severely mineralized. TDS has a Desired limit of 500 mg/l and a Permissible limit Table 2 Indian ratings of water quality index

WQI value

Rating of quality

Grade

1–25

Excellent

A

26–50

Good

B

51–75

Poor/bad

C

76–100

Very poor/very bad

D

>100

Unfit for drinking

E

332

R. Deopa and K. K. Singh

of 2000 mg/l, both of which are prescribed for drinking. TDS concentrations value from 747 mg/L to 1766 mg/L in this investigation. pH: There has been no health-based pH recommended value proposed. Although pH has little direct impact on consumers, it is one of the most important operational water quality factors, with the optimal pH requiring a range of 6.5 to 9.5 according to WHO. According to IS:10,500 (2012) [6], the acceptable pH range for drinking water is 6.5–8.5 for household use and living organisms. It is defined as the negative logarithm of the hydrogen ion concentration and is one of the most essential water quality measures. At 25 °C, pure water has a pH of approximately 7.0. The minimum and maximum values of pH were measured as 6.83 and 7.41, respectively. There was no well in which the pH exceeds 8.5 as given by IS 10500(2012). Alkalinity: The total of all titratable bases is the alkalinity of water, which is its acid neutralizing capacity. The presence of hydroxide ions, bicarbonate ions, and carbonate ions are the main causes. The permitted limit for hardness is 200 mg/L, and the reason for rejection is 600 mg/L, according to IS:10,500 (2012). The hardness value in our study region ranges from 65 to 110 mg/L. Hardness: It’s a term that describes the characteristics of highly mineralized waters. It is the concentration of multivalent metallic cations in solution. It is divided into two parts: carbonate hardness and non-carbonate hardness. The permitted limit for hardness is 200 mg/L, and the reason for rejection is 600 mg/L, according to IS:10,500 (2012). The value of hardness in our study region ranges from 125 mg/L to 411 mg/L. Chlorides: Chlorides can be found in different quantities in all natural waters. As the mineral concentration rises, the chloride content rises as well. Unless the water is brackish or saline, the chloride ion is present in natural waterways in relatively low amounts, usually less than 100 mg/L. Water and beverages with high chloride concentrations have a salty taste (World Health Organization). Chloride concentrations greater than 250 mg/L, on the other hand, can cause a perceptible taste in water. The chloride content of water samples in our study area is quite low, ranging from 7.5 to 35 mg/L (Fig. 3, Tables 3, 4, 5, 6, 7, 8 and 9). Among all physicochemical parameters Electroconductivity and Total Dissolved Solids have their values beyond acceptable limits for a particular time at that particular day. Hence, the relation between physicochemical parameters and the length is determined using image overlapping. The variation with distance is taken diagonally towards the centre. The results obtained are shown below (Figs. 4, 5 and 6 and Tables 10 and 11).

5 Conclusion On the basis of the finding, it was concluded that the water quality index of NIT Kurukshetra Campus is found to be Excellent. Although at every sampling location the Electroconductivity, Total Dissolved Solid (TDS) value is beyond the acceptable limit on the particular collected sample for that particular day although more samples

Influence on Water Characteristics Away from Various Sources of NIT …

333

Fig. 3 Spatial distribution maps for various physicochemical parameters

need to be collected for more concrete information but still the water quality index is found as excellent. From the various spatial distribution maps which we got through Kriging Technique of Interpolation by ArcGIS Software, we found that in case of Electroconductivity, Total Hardness and Total Dissolved solids, it is increasing beyond acceptable limits as we are moving away from the sources of water and getting maximum at the center of the NIT Kurukshetra campus. While Alkalinity, Chloride and pH values are well within permissible limits as suggested by the Indian Standard Code. It is found that Alkalinity is low at the Eastward side of the campus near Tubewell No.8, Tubewell No.6 and Tubewell No. 9 but as we move toward the westward side alkalinity is increasing at Tubewell No.7. The pH was found lowest at Tubewell No. 8 which is located at North-Eastward end of the campus and highest at southern source which is Tubewell No. 9 (at Harihar Bhawan). Chlorides values are found as lowest

334

R. Deopa and K. K. Singh

Table 3 Physico-chemical parameters of waters at location L1 , L2 , L3 S.No

Parameters

Acceptable limits (As per IS 10500 (2012)

Loc L1

Loc L2

Loc L3

1

Color

–

Clear

Clear

Clear

2

Odor

–

Agreeable

Agreeable

Agreeable

3

Taste

–

Agreeable

Agreeable

Agreeable

4

Salinity

–

1.85

1

0.86

5

Electroconductivity

1500 micromho/cm (WHO)

3548

1975

1727

6

TDS

500 mg/l

1766

983

859

7

Total Alkalinity

200 mg/l

83

110

71

8

Total Hardness

200 mg/l

411

150

139

9

pH

6.5–8.5

6.92

7.27

7.41

10

Chlorides

250 mg/l

20.2

35

7.5

Table 4 Physico-chemical parameters of waters at site L3 and L5 S.No

Parameters

Acceptable limits (As per IS 10500 (2012)

Loc L4

Loc L5

1

Color

–

Clear

Clear

2

Odor

–

Agreeable

Agreeable

3

Taste

–

Agreeable

Agreeable

4

Salinity

–

0.65

0.86

5

Electroconductivity

1500 micromho/cm (WHO)

1373

1692

6

TDS

500 mg/l

747

847

7

Total Alkalinity

200 mg/l

80

65

8

Total Hardness

200 mg/l

135

125

9

pH

6.5–8.5

6.97

6.83

10

Chlorides

250 mg/l

20

22

Table 5 Water quality index at location L1 (Tubewell No.6) S.No

Parameter

Standard value (Sn)

Observed value (Vn)

Sub-index of quality parameter (qn)

Weightage factor (W)

Water quality index

1

Electroconductivity

1500

3548

236.53

0.00496350365

20.34

2

Total dissolved solids

500

1766

353.20

0.01489051095

3

Total hardness

200

411

205.50

0.03722627737

4

Total alkalinity

200

83

41.50

0.03722627737

5

pH

8.5

6.92

5.33

0.8759124088

6

Chlorides

250

20.2

8.08

0.0297810219

Influence on Water Characteristics Away from Various Sources of NIT …

335

Table 6 Water quality index at location L2 (Tubewell No.6) S.No

Parameter

Standard value (Sn)

Observed value (Vn)

Sub-index of quality parameter (qn)

Weightage factor (W)

Water quality index 18.69

1

Electroconductivity

1500

1975

131.67

0.00496350365

2

Total dissolved solids

500

983

196.60

0.01489051095

3

Total hardness

200

150

75.00

0.03722627737

4

Total alkalinity

200

150

75.00

0.03722627737

5

pH

8.5

7.27

18.00

0.8759124088

6

Chlorides

250

35

14.00

0.0297810219

Table 7 Water quality index at location L3 (Tubewell No.8) S.No

Parameter

Standard value (Sn)

Observed value (Vn)

Sub-index of quality parameter (qn)

Weightage factor (W)

Water quality index 31.06

1

Electroconductivity

1500

1727

115.13

0.00496350365

2

Total dissolved solids

500

859

171.73

0.01489051095

3

Total hardness

200

139

69.33

0.03722627737

4

Total alkalinity

200

71

35.33

0.03722627737

5

pH

8.5

7.41

27.33

0.8759124088

6

Chlorides

250

7.5

3.00

0.0297810219

Table 8 Water quality index at location L4 (Tubewell No.9) S.No

Parameter

Standard value (Sn)

Observed value (Vn)

Sub-index of quality parameter (qn)

Weightage factor (W)

Water quality index 8.67

1

Electroconductivity

1500

1373

91.53

0.00496350365

2

Total dissolved solids

500

747

149.40

0.01489051095

3

Total hardness

200

135

67.50

0.03722627737

4

Total Alkalinity

200

80

40.00

0.03722627737

5

pH

8.5

6.97

2.00

0.8759124088

6

Chlorides

250

20

8.00

0.0297810219

336

R. Deopa and K. K. Singh

Table 9 Water quality index at location L5 (academic section) S.No

Parameter

Standard value (Sn)

Observed value (Vn)

Sub-index of quality parameter (qn)

Weightage factor (W)

Water quality index 77.26

1

Electroconductivity

1500

1692

112.80

0.00496350365

2

Total dissolved solids

500

847

169.40

0.01489051095

3

Total hardness

200

125

62.50

0.03722627737

4

Total alkalinity

200

65

32.50

0.03722627737

80.35

0.8759124088

8.80

0.0297810219

5

pH

8.5

6.83

6

Chlorides

250

22

Fig. 4 Image overlapping

at Tubewell No.9 and its value increasing as we move away from this source and found highest at Tubewell No.7 at the Western side of the campus.

Influence on Water Characteristics Away from Various Sources of NIT …

337

Fig. 5 Plot showing variation between TDS and length from various sources towards center location point L5

Fig. 6 Plot showing variation between EC and Length from various sources towards center location point L5

338

R. Deopa and K. K. Singh

Table 10 Variation of TDS with length from various stations towards the centre L5 (in meters) TDS value (in mg/L)

Distance (L3–L5)

Distance (L2–L5)

Distance (L4–L5)

803.5

130.5

0

0

916.5

348

56.5

0

1029.5

460

153

80.5

1142

516

237

203

1254.5

597.5

315.5

286.5

1367.5

674

387

359.5

1480.5

726.5

445

422.5

1608

778

504.5

480

1720.5

828

568

531

Table 11 Variation of EC with length from various stations towards the centre L5 (in meters) EC value (in mg/L)

Distance (L3–L5)

Distance (L2–L5)

Distance (L4–L5)

1736.5

149.5

42.5

0

1977

332.5

134

56.5

2217.5

400.5

227.5

165

2458

454

312.5

272.5

2698.5

509.5

387

352

2939.5

567.5

457.5

413.5

3180

624.5

525

488.5

3420.5

685

581

551

References 1. Munavalli GR, Mohan Kumar MS (2005) Water quality parameter estimation in a distribution system under dynamic state. Water Res 39(18):4287–4298. https://doi.org/10.1016/j.wat res.2005.07.043 2. Kirubakaran M, Ashokraj C, Colins Johnny J, Anjali R (2016) Groundwater quality analysis using WQI and GIS techniques : a case study of Manavalakurichi in Kanyakumari District, Tamilnadu, India groundwater quality analysis using WQI and GIS techniques : a case study of Manavalakurichi in Kanyakumari district, T. Int J I 2(February):341–347 3. Chabuk A, Al-Madhlom Q, Al-Maliki A, Al-Ansari N, Hussain HM, Laue J (2020) Water quality assessment along Tigris River (Iraq) using water quality index (WQI) and GIS software. Arab J Geosci 13(14). https://doi.org/10.1007/s12517-020-05575-5 4. Dhanraj MR, Rajesha R (2017) Assessment of ground water quality in Deralakatte Belma Panchayath Mangalore Dakshina Kannada. Int J Eng Trends Technol 44(1):12–16. https://doi. org/10.14445/22315381/ijett-v44p203 5. Naik V, Naik D, Naik G, Moger K, Yallurkar S (2019) Characterization of ground water based on water quality index in Bhatkal, Uttara Kannada District, Karnataka. pp 2199–2206 6. BIS (2012) Indian standard drinking water specification (Second Revision). Bur Indian Stand IS 10500(May):1–11 [Online]. http://cgwb.gov.in/Documents/WQ-standards.pdf 7. Anesthesia N et al (2010) Guidelines for 2010

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS in Mountainous Terrain Dolonchapa Prabhakar, Anoop Kumar Shukla, Babar Javed, and Satyavati Shukla

Abstract Landslides are a devastating natural phenomenon that can cause significant damage to people, property, and infrastructure. Various physical, geological, climatic, and tectonic factors can contribute to landslides in different parts of the world. In addition to natural causes, improper construction and chaotic development can also lead to landslides and the destruction of property and lives. Landslide hazard zonation (LHZ) is a method of dividing a landmass into homogeneous zones and ranking. Landslides are a pervasive and destructive hazard in the Himalayan regions. While they cannot be completely eliminated, their effects can be minimized. This study focuses on the Garhwal Himalaya region to identify areas prone to landslides. Geographical Information System (GIS) was used to create a database, analyze, and generate output. LISS III images were used to create the land use land cover (LULC) map and Landsat satellite data was used to identify. This research paper aims to integrate various techniques and tools to generate a geospatial database for landslide hazard zone delineation in the study area. LISS III images were used for land use land cover (LULC) mapping and Landsat satellite data for lineaments identification. The Analytical Hierarchy Process (AHP) approach was used to determine the weights of the landslide influence parameters. ERDAS Imagine and ArcGIS software were used to integrate the input layers after assigning them suitable weights. The primary objective of this research paper is to combine the various existing methods and tools to create a Geo-spatial database for delineating landslide hazard zones in the study area. The resulting map’s susceptibility to landslides has been divided into five categories, namely very low, low, moderate, high, and very high. This LHZ map can be of great benefit to planners and designers in selecting the most appropriate route paths.

D. Prabhakar Department of Civil Engineering, Indian Institute of Technology, Roorkee, Uttarakhand, India A. K. Shukla (B) · B. Javed Manipal School of Architecture and Planning, Manipal Academy of Higher Education, Manipal, Karnataka, India e-mail: [email protected] S. Shukla Key Laboratory of Geospatial Informatics, College of Earth Sciences, Guilin University of Technology, Guilin, PR China © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_25

339

340

D. Prabhakar et al.

Keywords Analytical Hierarchy Process (AHP) · GIS · Satellite data

1 Introduction Himalayan ranges extend up to a length of 2500 km from the East to West with mean width along full longitudinal extension ranging from 110 to 400 km. The massive area covers the mountain ranges of Geological younger ages. These ranges contain some fantastic altitude gradients. The entire Himalayan region is inherently unstable. Every year, landslides occur in the area stretching from Jammu & Kashmir to the Northeastern states, including Himachal Pradesh, Uttarakhand, and Sikkim Himalayas, resulting in significant loss of life and property. Various regions of India get affected differently by landslides due to the characteristic, physiographic, geological, climatological, and tectonic conditions. In count to natural occurrence to landslides, improper arrangement and chaotic construction are also responsible for increased incidences of landslides and consequent damages of properties and human lives. However recurrent landslides occur due to different causative factors like topographic, lithological, structural features, natural calamities like earthquake, heavy rainfall, cloud bursts, human activities, and improper development activities without considering the depth of natural factors, unscientific formation cutting for road construction, and use of heavy high explosives in blasting that produces instability. This increases the failure of slope stability and triggers mass movement resulting in landslides. The brunt of landslides is mainly felt on national highways which get blocked during rainy season when more and more landslide areas get triggered. The landslide prompts demolition of roadways and progressing territories where escape for individuals and vehicles and supply of help material in routine is relatively incomprehensible. Besides a huge sum is also required for removal of debris, prevention of further landslide and planning activities increases work pressure over the Border Road Organizations (BRO), which is involved in construction, development and maintaining the roads in hilly terrain of the Himalayan region. Landslide chance alludes to the idea of harm liable to be caused if failure happens. The harm caused might be in the form of death toll and wounds or potential loss of land and property. The degree of harm is subject to existing areas that utilize examples of the region prone to be influenced and its populace. For e.g., a littler landslide in thickly populated regions may cause broad harm as thought about a noteworthy avalanche in remote zones. Therefore, the risk due to landslide is a function of hazard probability and the damage potential [1]. With the application of satellite imagery and GIS data, combined with ground information, several qualitative and quantitative approaches have been used in different studies to assess landslide susceptibility. These include frequency ratio (FR) [2, 3], index of entropy (IOE) [4], fuzzy logic [5], weight of evidence (WOE) [6], logistic regression (LR) [7], analytical hierarchy process (AHP) [8], and a range of machine learning approaches. Even though there are various approaches for assessing landslide susceptibility, there is still no conclusion over the best approach in different regions. In the qualitative approaches AHP has been regarded as the best approach

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS …

341

to analyze landslide susceptibility, which is based on the knowledge and experts’ experiences. Many applications have been reported from several parts of the world [9–11]. Some works have also reported on the comparisons of the FR and AHP methods for landslide susceptibility assessment worldwide [12–18]. However, no study has been reported from the Parts of the Tehri and Pauri Garhwal districts in India’s Uttarakhand state in this aspect. Landslide hazards refer to the probability of occurrence of landslide danger. The probability of occurrence of landslide in a specific terrain can be determined, predicted, and mapped with density of distribution and the terrain can be divided in terms of hazard zones. The causative factors are natural and anthropogenic and their intensity triaging landslide can be categorized, and weight assigned to them can be considered as important factors. The zones of landslide density can be classified from Very Low Hazard (VLZ) to Very High Hazard (VHH). The identified Hazard zones can be considered for preliminary planning activity of development schemes and help to avoid unstable areas during the planning stage. Even though it may be unavoidable, recognizing the potential for landslides in the early stages of planning will help to develop better preventive measures as a priority. High hazard zones should be given the highest priority. In this regard, landslide hazard zonation mapping is an essential component of such a study. The landslide hazard zonation mapping in the vicinity of Rishikesh and along National Highway-58 can predict the zones of hazard in the region. This study will likewise be useful for doing the point-by-point mapping of the landslide risk appraisal mapping along the street which will feature danger of harm to the street and encompassing property, line of correspondences, schools, buildings, and neighborhoods independently. The research basically has focused on the landslide hazard zonation mapping around Rishikesh and the surrounding area which is covered within the Toposheet-53 J/8 of 1:50,000 scale of Survey of India. Validation of the result was carried out by landslide inventory mapping using GPS, Maps & other instruments along part of NH-58 from Rishikesh to Kaudiyala village that falls on the road 31 km from Rishikesh. The objectives of the study are to generate a Landslide Hazard Zonation map of the study area for effective and efficient disaster management in the future, and to determine the various parameters and importance of remote sensing and GIS in LHZ in the present study.

2 Objective of the Study Area The main aim of the study is to quantify the landslide Hazards around the surrounding area of Rishikesh and National Highway-58, by preparing the Landslide Hazard Zonation Map (LHZ) with the aid of remote sensing and GIS and validating the result by conducting a GPS field survey along a portion of the national highway that runs through the research area to map the inventory of landslides. The main objective of the research is as below.

342

D. Prabhakar et al.

• To create a detailed GIS database of different thematic parameters depending upon the degrees of influence on the landslide process. • To use GIS analysis techniques to evaluate the importance of the factors. • To prepare landslide hazard zonation map for the Rishikesh area. • Landslide hazard zonation map is to be prepared for part of National Highway-58 falling in the study area with a buffer zone of 2 km. • Preparation of landslide inventory map along NH-58 and validation of result carried out by using GPS field survey.

3 Description of Study Area and Datasets Used Parts of the Tehri and Pauri Garhwal districts in India’s Uttarakhand state make up the research area. The study area has a perimeter of approximately 100 km and a surface area of 625 km2 . The following Toposheet No: - 53 J/8 of Survey of India is covered. The area extent is bounded between longitude 78° 15' 00'' E and 78° 30' 00'' E and latitude 30° 00' 00'' N and 30°15 ' 00'' N (Fig. 1). Landslide Hazard Zonation Mapping utilizing Remote Sensing and GIS in the mountainous terrain involves the integration of various datasets to analyse factors contributing to landslide susceptibility. Some essential datasets used for this study are:

Fig. 1 Map showing location of the study area

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS …

343

• Satellite Imagery: Satellite images of LISS-III and Landsat series, provide valuable information about land cover, land use, and vegetation characteristics. • Digital Elevation Model (DEM): DEM data captures the terrain’s elevation and helps create slope, aspect, and curvature maps, which are essential for understanding the topographic influence on landslides. • Land Cover Land Use Maps (LULC): These maps categorize land cover types such as forests, urban areas, agricultural lands, and bare soil. They contribute to identifying areas prone to landslides due to land use changes. • Drainage Maps: Maps indicating drainage patterns, watershed boundaries, and proximity to water bodies help assess areas vulnerable to landslides due to erosion and saturation. • Road and Infrastructure Maps: Information on roads, highways, and other infrastructure aids in identifying areas where construction and human activities might influence landslide susceptibility. • Landslide Inventory Maps: Historical records of landslides, including their locations, types, and characteristics, contribute to understanding past events and validating hazard zonation results. These datasets, when integrated and analysed using remote sensing and GIS techniques, contribute to a comprehensive understanding of landslide susceptibility and facilitate the creation of accurate Landslide Hazard Zonation maps in the mountainous terrain.

4 Methodology The methodology adopted in this study are as follows: • Study Area Selection: Selection of area for study where landslide problems are in existence. • Data Collection and Compilation: Data availability /data procurement and collection. • Data Preprocessing: Selection of satellite images suitable for regional and local level landslide study for land use identification and image processing. • Collection of topographic maps relevant to satellite images and scale of study. • Collection of relevant secondary data from all sources and access previous studies conducted on the area. • Data organization, data pre-processing, import satellite data and Raster scan Topomap in GIS environment and provide projection system. • Generation of relevant thematic data like Contour map, DEM, Drainage map, Slope map, Aspect map, Drainage frequency map. • Lineament occurrence map, Lineament intersection map, Road map, Landslide inventory map and Location map. • Conversion of vector thematic to raster format using raster conversion tool.

344

D. Prabhakar et al.

• Weighting and Factor Ranking: Definition of landslide hazard classification and rating/weight for each thematic class as per their role in landslide triggering intensity in attributed information. • GIS analysis: Select GIS analysis tools for overlay operation by providing weight calculated for each thematic and prepare landslide zonation map. • Verification of the errors and consequent modification to appropriate levels. • Estimate geometrically, geographically, and statically landslide prone area with their hazard intensity class. • Verification of the land use land error map, geography of different hazard zones obtained by the GIS analysis methodology. • Map Generation and Visualization: Preparation of final thematic maps and landslide hazard zonation maps by reanalysis in GIS environment with symbolization. • Preparation of final print out maps at 1:50,000 or as needed scale.

4.1 GIS Analysis for Landslide Hazard Zonation Mapping Landslide Hazard Zonation Map is prepared in this work by using the weighted overlay method of raster analysis available in ArcGIS software. The Analytic Hierarchy Process (AHP) is used for the calculation of weight and rank assignment to different thematic parameters of landslide importance.

4.2 Analytic Hierarchy Process (AHP) Geographers and spatial organisers are typically intrigued by choice problems that involve a large number of plausible options as well as several evaluation criteria that are incompatible and conflicting. The options are typically examined by a variety of individuals (decision-makers, managers, subject matter experts, and interest groups), each of whom has their own unique preferences regarding the relative importance of the criteria used to evaluate the options. On the other hand, Spatial Multi-criteria Decision Analysis (SMCDA) or Multi-criteria Decision Making (MCDM) can be used to identify the best solution to a given problem by taking into account multiple criteria. By combining GIS and MCDM, it is possible to develop a more comprehensive approach to spatial planning and management. This approach can be used to identify the most suitable solution to a given problem, taking into account a range of geographical data and multiple criteria. Contrarily, MCDM and a wide range of related methodologies, including multi-attribute utility theory (MAUT), public choice theory, and collaborative decision-making, offer a collection of techniques and procedures to reveal decision makers’ preferences and to incorporate them into GISbased decision-making. To this purpose, a collection for studying geographic events can be referred to as GIS-based (or spatial) multi criteria decision analysis, where the outcomes of the analysis (decisions) depend on the spatial arrangement of the

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS …

345

events. Spatial analysis, as it is generally characterised, includes spatial multi-criteria analysis. Multi-criteria decision-making is not a well-established process or effectively integrated into the field of spatial analysis and GIS, despite the fact that the majority of spatial decision problems are multi-criteria in nature and involve economic, social, environmental, and political dimensions as well as conflicting values. The Spatial Multi-criteria Decision Analysis (SMCDA) is a kind of multi attribute decision, making the (MADM) process. The main elements of SMCDA are evaluation criteria, alternatives, and decision maker’s preferences. SMCDA issues ordinarily include criteria of changing significance to decision makers. Thus, data about the relative significance of the criteria is required. This is generally accomplished by doling out a weight to basis. The deduction of weights is a focal advance in inspiring the chief’s inclinations. A weight can be characterized as esteem allocated to an assessment foundation that shows its significance with respect to other rule in the general utility. In multi-criteria decision-making, a variety of criteria-weighing techniques have been developed. Some of the most common techniques are rating, ranking, trade-off analysis, and pairwise comparison. Authors in [19] created the pairwise comparison technique within the framework of the Analytic Hierarchy Process (AHP) as shown in Table 1. Pairwise correlations are used in this method to create a percentage network. It generates the relative weights as yield using pairwise examinations as input. The weights are specifically determined by normalising the eigen vector related to the extreme Eigen value of the (reciprocal) ratio matrix. The present work involved six parameters: slope, aspect, drainage frequency, lineament occurrences, Lineament intersections and land use. It necessitated determining their relative importance. This was accomplished by comparing each pair of parameters pairwise. There are three main steps in the process [20]. 1. Pairwise comparison-matrix development 2. Computation of the criterion weights 3. Estimation of the consistency ratio.

Table 1 Pairwise-comparison scale (Satty, 1980)

Significance of Issue

Description

1

Equal significance

2

Equal to moderate significance

3

Moderate significance

4

Moderate to strong significance

5

Strong significance

6

Strong to very strong significance

7

Very strong significance

8

Very to extremely strong significance

9

Extreme significance

346

D. Prabhakar et al.

A weighted evaluation system was employed in this investigation. The ArcGIS software’s built-in arithmetic weighted overlay method, which accepts both continuous and discrete grid layers and produces continuous grid data layers, was used to integrate all of the layers. The complete methodology is shown in the methodology section Fig. 2. Using the ArcGIS spatial analyst module, all the components are categorised according to the significance of each element impacting landslide hazard. Each subclass is assigned a rating between 0 and 9 in order of increasing risk; zero denotes little risk and 9 denotes extreme risk. All the rating classes are multiplied by the corresponding weight to obtain the landslide potential index map. Table 2 displays the weights and their ranking of the landslide influence elements. The Analytical Hierarchy process (AHP) is used to calculate the weights. High weights imply high landslides, whereas the lowest weights suggest low landslides. The resulting values are obtained the range from 32 to 331. It should be classified into five classes very low, low, moderate, high and very high, moving average with window size of 3, 7 and 9 were considered.

EXISTING DATA

REMOTE SENSING DATA

SRTM DEM

Drainage

Slope/ Aspect

Google Earth

LISS-III / LANDSAT

Pre-Processing

Field data (GPS)

Inventory map

Road MAP

Stream Map

Land cover

Fault

AHP

Map assessment

Validation and Landslide hazard zonation map generation

Fig. 2 Methodology flowchart for landslide hazard zonation mapping

Geology

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS …

347

Table 2 Computation of criterion weights Factor

Slope

Drainage frequency

Land use

Lineament intersection

Lineament occurrence

Aspect

Slope

1

2

2

2

8

9

Drainage frequency

1/2

1

2

3

6

9

Land use

1/2

1/2

1

2

9

6

Lineament Intersection

1/2

1/3

1/2

1

8

7

Lineament Occurrence

1/8

1/6

1/9

1/3

1

4

Aspect

1/9

1/9

1/7

1/6

1/4

1

The Analytical Hierarchy Process (AHP) offers a structured and systematic approach to decision-making, allowing for the effective prioritization of factors in complex problems. It enables the integration of both objective data and subjective inputs, enhancing the comprehensiveness and accuracy of analyses. However, AHP relies on expert judgments and pairwise comparisons, which can introduce subjectivity and bias. Its complexity in constructing hierarchies and assigning weights can be time-consuming, and it may struggle to handle uncertainties and sensitivity to weight allocation. Careful consideration of these pros and cons is essential when applying AHP in decision-making contexts.

5 Results and Discussions 5.1 Thematic Parameters and Their Relationship with Landslide Causes Slope Map Slope is a crucial factor in landslide research since its stability determines how frequently and how intensely landslides occur. According to [21], landslides occur most frequently when the slope angle is between 35° and 40°, and less frequently when the slope is more than 40°. According to reports, landslides occur more frequently in regions with slopes between 6° and 45° [22]. The Song River portion of Dehradun experienced numerous landslides that occurred at angles ranging from 18 to 45 degrees [23]. The Marapalam landslide in 1993, which occurred close to Coonoor, occurred on a slope of roughly 30° in the slope-forming material [24]. Slope angle was calculated and used in this study as a one theme layer for landslide analysis. The digitised contour map has been used to create the TIN (Triangulated irregular network) model for the research region. The Delaunay triangulation of

348

D. Prabhakar et al.

irregularly distributed nodes or observation points serves as the foundation for the TIN terrain model, which employs a sheet of continuous, connected triangular facets [25]. The TIN model was used to create the slope map. The majority of the research region, it was concluded, is sloped between 25° and 90°. There are four different types of slopes. The slope of 25° to 45° is more prone to landslides, according to observations made in the field using a handheld GPS and compass and from information accessible in a Geological Survey of India (GSI) article [26]. Therefore, this category is given a higher rank value. The largest frequency of landslides occurs on bare soils, crop and grass fallow areas, degraded forests, and barren land types as slopes rise. In locations with steep slopes, forests reduce the risk of landslides, but cultivation raised the risk. As the slope rises, vegetation clearing increased the likelihood of landslides. The requirement for a sufficiently thick mantle of residual soil prevents steep slopes from failing in reaction to rainfall. Rain is needed more to wet the bottom half of a slope where the weathered profile is dryer. Under the combined influences of comparable transpiration losses from a smaller moisture store and the increased gravitational pressure on the unsaturated down slope drainage, the steeper slope loses water more quickly. The roots close to the surface of the ground offer a component of strength adequate to withstand movement brought on by collapse of the deep residual soil. Aspect map The aspect map is yet another crucial element that affects the stability of the slope and ultimately landslides. It designates the slope’s orientation and relates to the potential exposure to light and shade that a mountain or hill slope may have. The terrain’s exposure to storm fronts is affected by the slope aspect. One of the main contributing reasons to a landslide is the slope’s aspect. Additionally, it has an impact on pore water pressure and alternation variations. Land use Land Cover (LULC) map The presence and movement of rainfall-induced landslides are significantly influenced by the vegetation cover. The behaviour of landslides is frequently altered by changes in plant cover [27]. It has been determined that via numerous examinations that land use and land cover, particularly of the woody variety with robust and substantial root systems, contribute to the improvement of slope stability [28]. As a result of changes in land use over time, it is possible that forest land has been converted to another land use, such as farmland. In southern Honduras, especially when the slope increased, deep-rooted vegetation that stabilises the topsoil, such as shrubbery or forest, was a very significant role in lessening landslide hazard [22]. The shear strength of the soils is increased, according to [29], by vegetation roots that saturate the entire surface. In comparison to regions with less or no vegetation, places with denser vegetation were thought to be less prone to sliding [30]. However, at Dehradun and Mussoorie, 91% of the landslides happened in regions that weren’t covered in trees, indicating the role that vegetation plays in the beginning of slope instability [31]. A land use land cover map on a scale of 1:50,000 was created using

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS …

349

LISS III images and supervised categorization. In this study area, land use is categorized into eight types. (1) Dense Vegetation (More than 40% Canopy), (2) Medium Vegetation (10 to 20% Canopy), (3) Degraded Vegetation (More than 40% Canopy), (4) Barren Land (Non-Rocky, wasteland), (5) Agriculture land, (6) Build up area, (7) River Sand and (8) Water Body. For validation purpose, land use and land cover classification accuracy assessment were performed which was found to be the Overall Accuracy which is 91.50 and Kappa Coefficient being 0.90. Map of Drainage Frequency In hilly places, drainage is a significant cause of landslides. Landslides that tend to occur less frequently are from the drainage network because shear strength is low and close to the drainage network due to the greater amount of water seepage there. Landslides occur more frequently close to the drainage network than farther from it in the study area. The initiation of a landslide may also be influenced by the terrain modification brought on by gully erosion [32]. A drainage map on a scale of 1: 50,000 was created from a topographic map in order to understand the function of drainage. The drainage map was superimposed over a 0.50 × 0.50 km grid cover. Each grid’s drainage lines are tallied in order to determine the drainage frequency value. Based on the quantity of drainage incidents, the study region has been divided into five types. Lineament incidence map Fracture, joints, faults, bedding, and other features that are represented by lineament. These structures’ influence encourages infiltration and the growth of hydrostatic pressure on the material that forms slopes [33]. Landslides and faults are closely related; 88% of all landslides were found within 250 m or less of major faults [34]. All the landslides in the Song River region happened on fault scarps on the escarpment side, with the formations descending into the hill [35]. By using visual interpretation, a lineament map has been created from IRS IC- LISS III satellite images (see Fig. 3a and b). The map was superimposed on a grid with a size of 0.50 × 0.50 km, and the number of lineaments found within each grid was counted to determine the lineament frequency. The final map was digitally transformed and digitised. Based on the density of lineaments, an order of importance has been determined, and ranks have been allocated in accordance. Lineament intersection map Each grid’s lineament intersection count was calculated to produce the lineament intersection map. This map has been transformed into a digital file. The likelihood of a landslide occurring is increased with more intersections. Thematic parameters and their relationship with landslide cause the following criterion generated in Table 3 was generated.

350

D. Prabhakar et al.

Fig. 3 a Showing lineament map with false colour composite (FCC) image along NH-58. b Showing lineament map natural colour composite image along NH-58

Landslide Hazard Zonation Mapping Using Remote Sensing and GIS …

351

Table 3 Thematic/parameters and rank values S.No

Thematic/Parameters

Categories

Rank

1

Slope map

0°–15°

2

16°–25°

7

26°–45°

9

>45°

3

2

3

4

5

6

Aspect

Land use

Lineament occurrences

Lineament intersection map

Drainage frequency map

Flat

1

North, NE facing

2

East, SE, NW facing

3

West facing

4

SW facing

5

South facing

6

Dense vegetation (More than 40% canopy)

0

Medium vegetation (10 to 20% canopy)

3

Degraded vegetation (More than 40% canopy)

6

Barren land (Non-Rocky, wasteland)

5

Agriculture land

6

Built up area

5

River sand

2

Waterbodies

0

0 (Number of lineament occurrence within 1 0 km2 ) 1–2

2

3–4

4

5&>

8

0 (Number of lineament intersection)

0

1–2

3

3–4

6

100

6

7

8

9

10

Slope (%)

Aspect

Distance from major roads (Meter)

Distance from build-up area (Meter)

Distance from PTL(Meter)

Classification

Rated value

Remark

AHP weight (%)

1

Least suitable

10

100–200

2

Less suitable

200–300

3

Moderately suitable

300–400

4

Suitable

>400

5

Most suitable

15

1

Least suitable

S

5

Most suitable

SE + SW

4

Suitable

E+W

3

Moderately suitable

NE + NW

2

Less suitable

N

1

Least suitable

20,000

1

Least suitable

20,000

1

Least suitable

20,000

1

Least suitable

8

8

12

8

14

Potential Zones Identification to Effectively Exploit Solar and Wind …

419

Table 2 Classification and AHP weightages for wind site suitability Criteria

Classification

Rated value

Remark

AHP weight (%)

Wind speed (m/s)

4.6

5

Extremely suitable

40

1

Less suitable

20,000

1

Less suitable

20,000

1

Less suitable

20,000

1

Less suitable

20

15

10

15

5 Results and Concluding Remarks The study concludes that 1341.50 sq.km of the total area is the most suitable for Solar energy harnessing. 34,591.90 sq.km of the total is Suitable for Solar Potential Harnessing while 20,964.32 sq.km is Moderately Suitable. Karimganj, Silchar and Jorhat are the districts where Solar Harvesting suitability (Most Suitable) regions have been found. For Wind Potential totally 666.73 sq.km of the area is extremely suitable and 35,888.08 sq.km of the area is highly suitable and. 21,290.54 sq.km is suitable. The

420

P. T. Raj and N. S. R. Prasad

suitability map resulted in suitability in fragments majorly over the North-western part of Assam. Extreme Suitability is found in the North western region of Assam and Northern part of Assam that include districts like Bongaigoan, Barpeta, Nalbari, Goalpara, Darrang, Northern part of Karbi-Anglong Region and other districts around them.

Fig. 4 Restricted area map for solar suitability

Fig. 5 Restricted area map for wind Suitability

Potential Zones Identification to Effectively Exploit Solar and Wind …

421

North Eastern states especially Assam have hilly terrains making transmissions difficult, therefore medium scale installations in the remotest parts will avoid the need of long-distance transmission and energy losses. Bringing electricity to backward and tribal regions becomes an important aspect towards development. The basic necessity to find the suitable sites is fulfilled by this study by considering all the factors that impact the productivity and output of the renewable systems. Giving highest weightages to the ones that have high priority is made possible using the GIS and AHP model. With increased energy consumption and demand, finding an alternative source of energy is a necessity. Indiscriminate use of non-renewable sources of energy not only leads to exhaustion of the resource but also brings the world closer to global warming. While there are numerous villages and households in our country having no electricity there are also frequent fluctuations and electricity cuts.

Fig. 6 Reclassified criteria maps for solar suitability

422

P. T. Raj and N. S. R. Prasad

Fig. 7 Reclassified criteria maps for solar suitability

Large scale and medium scale solar plantation along with wind turbine installations would create a hybrid approach to mitigate the electricity crisis being faced by the country. This study gave a comprehensive understanding of the conditions that solar and wind energy sources demand for effective working at the same time ensuring the non-feasible areas are excluded from suitability. The study is oriented at the MCDA integrated approach by taking into consideration the environmental, economic and technical limitations making it a multi-aspect study. The technical and economic limitations are also included making it desirable in terms of planning and execution. Therefore, this study can help the energy planners take decisions (Figs. 4, 5, 6, 7, 8, 9, 10, 11 and 12).

Potential Zones Identification to Effectively Exploit Solar and Wind …

Fig. 8 Reclassified criteria maps for wind suitability

Fig. 9 Result: potential zones identified for solar site suitability in Assam

423

424

P. T. Raj and N. S. R. Prasad

Fig. 10 Area covered by solar suitability classes in percentage

Fig. 11 Result: potential zones identified for wind site suitability in Assam

Fig. 12 Area covered by wind site suitability classes in percentage

Potential Zones Identification to Effectively Exploit Solar and Wind …

425

References 1. Baseer MA, Rehman S, Meyer JP, Alam MdM. https://www.researchgate.net/publication/320 328038_GIS-based_site_suitability_analysis_for_wind_farm_development_in_Saudi_Arabia 2. Tomaszewski JM, Moriarty JK. https://www.researchgate.net/publication/328707431_Do_ wind_turbines_pose_roll_hazards_to_light_aircraft 3. Langston RHW, Pullan JD. https://tethys.pnnl.gov/sites/default/files/publications/Langston% 20and%20Pullan%202003.pdf 4. Kumar S, Samsoor Ali MS, Hameed A, Arun PR. https://www.researchgate.net/publication/259 390231_Impact_of_wind_turbines_on_birds_a_case_study_from_Gujarat_India

Prediction of Soil Organic Carbon in Unscientific Coal Mining Area Using Landsat Auxiliary Data Naorem Janaki Singh, Lala I. P. Ray, Sanjay-Swami, and A. K. Singh

Abstract Unscientific coal mining is affecting soil attributes, deteriorating soil health and crop productivity, and is reflected by the soil organic carbon (SOC) content. The quantification of SOC is challenging with limited resource availability; however, satellite covariance is the alternate source of SOC determination with minimum labour requirement and limited laboratory facilities. An attempt is made to estimate the SOC using Landsat data and a model is developed by evaluating stepwise and enter/removal regression approaches. Fourteen predictor variables were used to build models and evaluate the prediction accuracy. Results showed that the SOC ranges 0.81–2.41% under unscientific coal mine affected sites; NDVI and BSI range 0.16– 0.61 and −0.35–0.12 with mean 0.32 and −0.03, respectively. SOC is correlated with RI (r −0.33) and GRVI (r 0.34). The enter (all variables in a block enter in a single step) approach linear regression model (Model 3) using multiple variables can explain only 43% SOC (RMSE 0.66 and R2 0.43). The stepwise linear regression model (Model 2) and Model 3 predicted the SOC as 6.94 and 3.78% higher than the actual SOC data. The model performance is increased by using multiple variables which may subside the less number of soil samples. Keywords Unscientific coal mine · Soil organic carbon · Indices

1 Introduction India is the fifth largest coal reserve and fourth largest coal producer (730.87 million tons in 2019–2020) in the world [1]. The coal mining in Meghalaya, NE India, was started on a small scale in the nineteenth century during the British period. The active mining was begun in the mid-1970s. Large-scale unscientific coal mining (such as rat-hole or open mining) was begun since 1980s in Jaintia hills without the permission of environmental clearance and safety measures of mining workers [2]. The Coal Mining Act, 1973, was not followed in Meghalaya due to the 6th schedule of N. J. Singh (B) · L. I. P. Ray · Sanjay-Swami · A. K. Singh College of PG Studies in Agricultural Sciences, CAU (Imphal), Umiam, Meghalaya 793103, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 S. Mesapam et al. (eds.), Developments and Applications of Geomatics, Lecture Notes in Civil Engineering 450, https://doi.org/10.1007/978-981-99-8568-5_31

427

428

N. J. Singh et al.

Constitution of India (state where the landowners were the owners of the minerals below their land) and Article 371 (which allowed state government to formulate its own coal policy in order to recognize customary tribal laws). The increased mining activity from 3.26% in 1975 to 10.75% in 2001 caused a decease in the forest cover from 22.5% in 1975 to 12.34% in 2001 and also crop area from 2.65% in 1975 to 1.62% in 2001 [3]. The mine spoils such as gravels, rocks, sand, and soil were piled and dumped causing damage to the surrounding vegetation, every million ton of coal extraction could damage about 4 ha of surface area [4]. The open cast mining extracted about 250 million tons and about 100 million tons of mine spoil are overburden the nearby land as a dumping site. Subsequently, the acid mine drain from such dumping site affect the soil properties of agricultural field [5, 6]. Most of the agricultural lands were turned into the meadow (grass land). The quantification of key component of soil health—soil organic carbon (SOC) is urgently necessary for alternative suggestions for improving the livelihood of the society. Many authors attempted to study soil organic carbon (SOC) using equations derived from soil properties and related environmental variables derived from remote sensing and geographical database [7]. Mulder et al. [8] digitally mapped the SOC using the spectral signature of soil which is related to soil colour depending upon humic acid content and its composition [9]. Digital mapping covers larger areas and also overcome the limitations of laboratory facilities and intensive sampling with advantage of satellite data [10]. The spectral mainly used for SOC prediction are 450, 590 and 664 nm at a visible range, and 1600–1900 nm, 2100 nm and 2300 nm short-wave infrared (SWIR) [11, 12]. Various vegetation indices were derived to study the SOC [13, 14]; selection of the most suitable vegetation index in SOC prediction is a major challenge. In this study, it is aimed to assess the SOC accumulation pattern using various indices derived from Landsat data with objectives (i) to find out the model promising predictor variables and (ii) to construct the suitable regression model for mapping the SOC. The high-resolution data has limitation over the coverage of areas. The open cast and rat-hole coal mining is carried out in wide scattering throughout the Jaintia Hills district, Meghalaya. The multiple vegetation indices could be derived from the multi-spectral sensor data such as blue, green, red, near infrared (NIR), short-wave infrared (SWIR), and thermal band. These indices are the input variables (independent variable) for prediction of dependent variable soil organic carbon (SOC) in the statistical approach. A medium resolution, large area coverage with multiple spectral band satellite data could be the best option for achieving such study. The constraint in the ground sample collection may be partially overcome with the help of space data which is an alternate source of soil attribute analysis for the well-being of farmer community.

Prediction of Soil Organic Carbon in Unscientific Coal Mining Area …

429

2 Material and Methodology 2.1 Study Area Jaintia hills covers an area of 3,819 km2 and lies between longitude 91° 51’–92° 45’ E and latitude 25° 5’–25° 4’ N with elevation 1627 m above sea level. The average annual rainfall is 1326 mm and temperature ranged from 7.8 to 24.5 ° C. About 40 million tons of coal (thin layer (30–212 cm), sub-bituminous type with high sulphur content (2.7–5.0%), low ash content (1.3–24.7%), moisture content (0.4–9.2%), high volatile matter and high calorific value (5,694–8230 kilo calories/kg)) is reserved [2, 15]. The agricultural fields were turned into the meadow with the effect of mining activity, and crops in the lowland area are limited to paddy. NBBSLUP (1996) published soil map (1:500,000), physiographic map (1:1600, 000), agro-ecological region (1:1600, 000), and geology (1:1600, 000) are available for this area.

2.2 Soil Sampling and Process The random ten (10) soil samples were collected at 0–20 cm from each soil mapping unit (SMU) occurring in a common area obtained after intersecting soil properties, climate, geology and physiography [16] in QGIS 2.18 on the basis of purposive sampling method [17] and the SCORPAN model [18]. The coordinate of each sample point was recorded using a Global Positioning System (GPS) with an accuracy of ±3 m. One composite sample was made from each ten random samples and 114 total composite samples are collected (Figs. 1 and 2). The average of ten coordinates represented the location coordinate of a composite sample. The soil samples were air-dried in the laboratory and sieved through 1 mm for soil organic carbon analysis by wet oxidized method (Walkley and Black, 1934). The SCORPAN model [18] is giving below Sa = f (s, c, o, r, p, a, n) where Sa is soil properties, f is the relationship function between soil conditions and the covariates, s is the information of soil, c is the climate at the site, o is an organism (vegetation, fauna or human activities), r is topography, p is parent materials, a is age and n is space. The sampling sites occurred in two climatic zones (i.e. warm prehumid thermic and warm prehumid hyperthermic) and two physiographic units (i.e. intermediate plateau and lower plateau) and geology (Jaintia series and Disang series (ultrabasic in deep shades)). The SMU of sampling sites are described below

430

N. J. Singh et al.

Fig. 1 The soil sampling sites from the common area obtained after intersection of soil map, climate, geology, and physiography. (Source https://megsoil.gov.in/images/meghalaya_soil_map. jpg)

2.2.1

Soils of Upper Plateau

SMU 01—Soils on moderately sloping are loamy texture, moderate erosion and deep and excessively drained, which is associated with coarse loamy, moderately deep and excessive drained, very severe erosion and strong stoniness on gently sloping. The taxonomy is Typic Kandiudults (covering 174900 ha). SMU 02—Soils on gently sloping are loamy texture, moderate erosion, deep and excessively drained which is associated with loamy soil, very slight erosion, ground water level